(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Impact of advertising on consumer price sensitivity : a behavioral analysis"

THE IMPACT OF ADVERTISING ON CONSUMER PRICE SENSITIVITY: 

A BEHAVIORAL ANALYSIS 



By 

ANUSREE MITRA 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL 

OF THE UNIVERSITY OF FLORIDA ON PARTIAL FULFILLMENT 

OF THE THE REQUIREMENTS FOR THE DEGREE OF 

DOCTOR OF PHILOSOPHY 



UNIVERSITY OF FLORIDA 



1990 



In loving memory of my mother — my strongest supporter, 
best friend and constant source of wisdom and strength, who 
always inspired me to scale greater heights. 



ACKNOWLEDGMENTS 

It is true — there really is light at the end of the 
tunnel! I'm not sure that I fully realize the implication of 
the statement as yet. But what I do feel is appreciation for 
all those who have helped make this dissertation a reality. 

I wish to offer my deepest thanks and heartfelt 
gratitude to the person without whom this dissertation would 
have always remained a dream--my advisor and chairman. 
Professor John Lynch. At times when everything else seemed 
to be going wrong, when life posed problems that made 
academic targets seem quite meaningless, his constant support 
was all I had. He has been beside me on this project from 
start to finish and has provided insightful comments and 
careful critiques, all of which significantly influenced the 
conceptual analysis and research methodology. Like a true 
friend he has has been jubilant at my successes and has urged 
me on when my spirits were flagging. John has taught me, by 
his own example, that striving for excellence can exist side 
by side with compassion and empathy for people. 

I am grateful to Professors Joe Alba and Wes Hutchinson 
for their encouragement and guidance. Their suggestions 
provided direction for making refinements in the study 
design. Thanks are also offered to Professor Bart Weitz for 
helping position this work and to Professor Rich Romano for 
sharing insights about the economics literature. 



Ill 



I would like to thank Nielsen-Cadbury Canada, Rowntree 
Macintosh Canada and Hershey Canada for making available the 
television commercials and also for generously providing 
samples of the candy bars that were used for the experiment. 

Very special thanks go out to Eric Olson, for writing 
the computer program that was used for the study. In spite 
of his busy schedule, he has made himself available and 
forgave all the impositions that I made on his time. I 
shudder to think of how much longer the data collection and 
analysis would have taken, had it not been for him. Eric 
entertained all my requests--small and big, and as a poor 
graduate student I could not offer any monetary incentive 
suitable for the amount and quality of effort that he put in. 
Although in such instances words are quite inadequate to 
express the depth of feeling, for the lack of a better 
option, it is with words that I must offer my sincerest 
thanks to him for making this dissertation possible within a 
reasonable length of time. 

The past few years have been hard, but I was also lucky 
to have a few good friends whose faith in me has been a 
reservoir of strength that I could draw on when everything 
else seemed hopeless. I wish to express heartfelt thanks and 
gratitude to Malay and Dola Ghosh, Sudip and Sarmila, and all 
my friends in the Marketing Department, who in their 
respective ways, have helped me through the rough times. 

Finally, I would like to express my sincere thanks to my 

family for their constant support and unquestioning faith in 

iv 



me. At times when the future seemed uncertain and bleak, 
what kept me going was the vision of their happy faces on 
seeing me realize a target that I had ventured out to 
achieve, a long time ago. 



V 



TABLE OF CONTENTS 

ACKNOWLEDGMENTS iii 

ABSTRACT viii 

CHAPTERS 

1 INTRODUCTION 1 

Effects of Advertising on Price Elasticity... 3 

A Behavioral Approach 4 

Objective 5 



2 TOWARD A CONCEPTUAL FRAMEWORK FOR EXAMINING 
THE IMPACT OF ADVERTISING ON CONSUMER PRICE 
SENSITIVITY 8 

Chapter Overview 8 

Economic Theoretic Views on the Impact of 

Advertising 8 

Impact of Advertising on Consumer Price 

Sensitivity 12 

Conceptual Framework 18 



3 AN EMPIRICAL INVESTIGATION OF THE IMPACT OF 
ADVERTISING ON PRICE ELASTICITY: CONCEPTUAL 
HYPOTHESES AND STUDY DESIGN 32 

Objective 32 

Background 32 

Hypotheses 4 6 

Overview of Research Methodology 47 



4 PRETESTS 51 

Overview 51 

Experiment 1 52 

Experiment 2 68 



VI 



5 AN EMPIRICAL INVESTIGATION OF THE IMPACT OF 
ADVERTISING ON PRICE ELASTICITY: RESEARCH 
METHODOLOGY 73 

Chapter Overview 73 

Subjects 74 

Stimuli 75 

Experimental Factors 76 

Dependent Variables 78 

Stimulus Design 7 9 

Procedure 80 



6 AN EMPIRICAL INVESTIGATION OF THE IMPACT OF 

ADVERTISING ON PRICE ELASTICITY: RESULTS 8 9 

Overview 8 9 

Results 89 



7 GENERAL DISCUSSION AND IMPLICATIONS 119 

APPENDIX 139 

Introduction and Preliminary Instructions... 140 
Price Profiles for the Twelve Brands 

Across First Eight Shopping Trips 158 

REFERENCES 177 

BIOGRAPHICAL SKETCH 183 



Vll 



Abstract of Dissertation Presented to the Graduate School 

of the University of Florida in Partial Fulfillment of the 

Requirements for the Degree of Doctor of Philosophy 



THE IMPACT OF ADVERTISING ON CONSUMER PRICE SENSITIVITY; 

A BEHAVIORAL ANALYSIS 

By 

Anusree Mitra 
August 1990 



Chairman: John G. Lynch, Jr. 
Major Department: Marketing 

The dissertation investigates the impact of advertising 
on consumer price sensitivity. Two opposing schools of 
economic thought, the information school and the market power 
school, have postulated divergent theoretical accounts about 
the way advertising influences consumer behavior and have 
made conflicting predictions about the direction of change in 
price sensitivity. The fragmentary empirical evidence does 
not provide clear support for either theory. 

This dissertation postulates that two basic theoretical 

constructs 'mediate the effect of advertising on price 

sensitivity: size of the consideration set and the dispersion 

of brand utilities. It is suggested that advertising content 

and the decision environment have a potentially different 

effects on these two constructs. 

viii 



Three experiments are conducted to examine the role of 
such factors in determining the relationship between 
\/ advertising and price elasticity. The first two studies 
examine the effects of advertising on the mediating 
constructs, and provide evidence that advertising could have 
an effect on the memory for brand names and on the relative 
preference for the brand alternatives. 

The third experiment uses a computer based shopping 
simulation to test 'the relationship between advertising and 
price elasticity, mediated by the two key constructs. 
Subjects went on shopping trips under two types of decision 
environments and three advertising conditions. The effects 
on the size of the consideration set, the dispersion of brand 
preferences and price elasticity were examined. 

The results indicate a very strong interaction effect of 
advertising and decision environment on the size of the 
consideration set. The content of advertising also had a 
significant impact on revealed preference for the various 
brands. The direction of change in price elasticity was as 
predicted on the basis of the hypothesized mediating roles of 
the basic theoretical constructs. 

The experiments provide an improved understanding of the 
relationship between advertising and price elasticity, by f 
isolating the underlying causal mechanisms . The results I 

highlight the important mediating influence of consideration 

f 
set size and dispersion of brand-utilities m predicting when i 

f 

advertising will have a particular effect. I 

ix I 



CHAPTER I 
INTRODUCTION 

In today's competitive environment, consumers are 
exposed to a significant amount of mass media advertising on 
a regular basis. Each day brings with it new exposures to 
this powerful method of marketer-controlled communication. 
Because advertising is such an inescapable part of a 
consumer's life, consumer researchers have paid considerable 
attention to the study of consumer responses to this element 
of the marketing mix. The marketing practitioner, too, is 
interested in consumer responses to advertising, which 
represents a strategic management tool that the firm can 
employ to achieve its goals of maximizing profits or market 
share. Advertising expenditures comprise a substantial 
portion of the firm's promotion budget; hence the marketer is 
interested in the long-term consequences of advertising 
versus other methods of building customer loyalty (Moran 
1978) . 

Aside from the firms and the consumers who are directly 
involved with its consequences, economists and regulators 
also study the effects of advertising. The economist, 
primarily interested in the interactions between the behavior 
of the firm and the behavior of the consumer, mainly focuses 
on the determination of the equilibrium level of advertising, 
its effect on total consumption, industry profits, industrial 



concentration and price levels (Albion & Farris 1981; Norris 

1984) . Economists are not concerned with consumer responses 

to advertising per se, but mainly with its implications from 

a social welfare perspective. The regulatory agencies are 

charged with the responsibility of providing a fair and 

competitive environment in which firms may operate, and of 

ensuring that marketers' activities do not prove to be 

detrimental to consumer welfare. Thus, the economic effect 

of advertising is also of interest to the public policy maker 

in charge of advertising regulation and consumer protection 

(Bloom 1976) . 

Even though they are studying the same phenomena, 

specialists in one field often have little awareness for the 

activities of researchers in closely related fields. But 

there remain a number of valid reasons why the interaction 

between disciplines is important. In many cases, the output 

of one field of study serves as the input of another. Also, 

the tools and techniques of one discipline can be effective 

in answering questions previously studied only by a different 

discipline. Finally, good policy decisions can benefit from 

I 
the knowledge and insights that different disciplines have to | 

offer. Thus, our understanding of the economic effects of 

advertising may be greatly enhanced by the adoption of an 

interdisciplinary approach. 

The focus of this dissertation is on the economic effect 

of advertising on consumer behavior. This is a topic that 



has received considerable attention in industrial 
organization and social welfare economics. Although the 
research on the economic effects of advertising originated in 
the field of economics, its conclusions have been the subject 
of a long-standing controversy. Economic theory does not 
make any reliable predictions regarding the effects of 
advertising. The pragmatic implication of this lack of 
theoretical foundation is that none of these controversial 
issues can be settled by a priori reasoning. This has 
resulted in unresolved disputes about whether advertising 
raises or lowers profits, leads to increases or decreases in 
prices paid, increases or decreases industry competition, and 
the like, with sweeping generalizations made on particular 
samples of data (Albion & Farris 1981) . 

Effects of Advertising on Price Elasticity 
Although much of the prior research in this area has 
discussed this issue in the context of industry competition, 
this dissertation will examine the effect of advertising on 
consumer price sensitivity. Economists have provided 
divergent theoretical accounts and have made conflicting 
predictions about the economic effects of advertising based 
on implicit assumptions about consumer responses to the 
information provided by advertising. The empirical research 
that has appeared in the economics and marketing literature 
has only tested deductions from these theories and has not 



been able to offer any conclusive evidence on this 
controversial issue. 

In this past research, advertising is conceptualized in 
terms of the total expenditure on this element of the 
marketing mix. None of these studies has explicitly taken 
into account variations in advertising content. It is 
suggested in this dissertation that the results could vary 
with advertising content, because consumer responses to 
different types of advertising messages are likely to differ. 
It is the central thesis of this paper that assertions about 
the potential effects of advertising on the consumer require 
a more detailed analysis of the consumer's decision process. 

A Behavioral Approach 
This dissertation adopts a behavioral approach in an 
attempt to arrive at an improved understanding of the impact 
of advertising on consumer price sensitivity. It draws on 
insights from behavioral literature and focuses on the 
consumer's choice decision with the objective of identifying 
the various possible mechanisms by which advertising might 
affect price sensitivity. In the conceptual framework that 
is adopted, it is suggested that the effect of advertising on 
price elasticity could be better understood by examining how 
advertising influences two mediating variables: the size of 
the consideration set and the dispersion of brand-utilities. 



Advertising-induced changes in the size of the 
consideration set . The consumer's 'consideration' set 
consists of those brands that the consumer actively considers 
at the time of choice. It is proposed that advertising might 
have a role in placing a brand in the consumer's 
consideration set and/or keeping other brands out of the set. 
In other words, advertising might influence consumer choice 

by ensuring that the sponsored brand is considered at the [ 

[ 
time of decision making and by simultaneously determining the 

f 
other alternatives that will be considered along with the I 

l; 

if 

advertised brand. f 

i, 

Advertising-induced changes on interbrand dispersion in j 

utilities. Another way in which advertising for a particular [" 

I 
brand might affect choice is by changing the dispersion of 

utilities for the various alternatives. Thus if advertising 

results in some brands being perceived to be clearly superior 

to competitive offerings, these brands are likely to be i 

relatively insulated from price competition. These factors ! 

I 
are likely to have implications for price sensitivity. [ 

I 

t 
Objective i. 

t 
This dissertation attempts to demonstrate that prior t 

\ 
conceptualizations that have addressed only one possible I 

[ 
relationship between advertising and price sensitivity are ^ 

too simplistic. In other words, there are different causal 

mechanisms by which advertising influences consumer price 



6 



sensitivity. The objective of this paper is to provide 
insights on the mechanisms by which advertising might affect 
consumer behavior and to identify conditions under which 
these are likely to hold. Using price elasticity as the 
dependent measure, several propositions that follow from 
behavioral assumptions about consumer responses to 
advertising are tested in a laboratory setting. 

Chapter II begins with a survey of the literature on the 
economic effects of advertising. Relevant theoretical and 
empirical research on the effect of advertising on price 
elasticity is reviewed. The limitations of previous research 
in making any theoretical statements about the effects of 
advertising on consumer behavior are discussed. A conceptual 
framework is proposed that attempts to better understand and 
predict the effects of advertising on consumer price 
sensitivity . 

In Chapter III the specific focus of the dissertation is 
articulated. An experiment designed to test several research 
hypotheses on advertising's economic effects is presented. 
Chapter IV presents some propositions, methods and results of 
two experiments that were designed to serve as pretests for 
the Experiment 3, which examined the effects of advertising 
on price elasticity. Chapter V presents the research 
methodology employed to test the hypotheses. Chapter VI 
presents the results of the third experiment, which provide 
strong support for the primary hypotheses. 



Chapter VII provides a general discussion of the various 
findings and the implications of these results for research 
on the effects of advertising on consumer behavior. 
Managerial and public policy implications of these findings 
and directions for future research are also highlighted. 



CHAPTER II 

TOWARD A CONCEPTUAL FRAMEWORK FOR EXAMINING THE IMPACT OF 
ADVERTISING ON CONSUMER PRICE SENSITIVITY 



Chapter Overview 
Historically, the effect of advertising on price 
sensitivity has been researched by economists. This chapter 
reviews the relevant background literature on this topic. 
Two major economic theories of advertising's effects on price 
elasticity are presented along with empirical evidence. The 
next section details the conceptual framework that is 
proposed to examine the economic effect of advertising on 
consumer behavior. 

Economic Theoretic Views of the Role of Advertising 
The dominant paradigm in economics assumes that each 
economic unit (the firm or the consumer) acts as if it is 
solving a constrained maximization problem. The simultaneous 
solution of these problems for every economic unit is called 
a 'market equilibrium. ' The most Important example of this 
paradigm is the general equilibrium model of perfect 
competition in which there are two types of economic units — 
consumers and firms. Certain characteristics of the 
consumers' preferences or 'tastes' are taken as axioms of the 
model and consumer tastes may be mathematically represented 
by a utility function. It is assumed that the consumer is 



rational and chooses quantities of various goods to cons 



ume 



m an attempt to maximize utility. In such a model, prices 
are determined through the market equilibrium of demand for 
goods by consumers and their supply by firms. 

Economic price theory did not include advertising until 
the seminal work of Chamberlain (1933) . Since that time, it 
has been theorized in normative models that there is an 
optimal level of advertising, which is inversely related to 
price elasticity (Dorfman & Steiner 1954; Nerlove & Arrow 
1962) . However, there does not exist a generally accepted 
theoretical view about the economic consequences of 
advertising. Given this lack of any generally accepted 
theoretical formulation, divergent streams of research have 
appeared in the literature. Thus, at the risk of 
oversimplifying the subtleties of individual economist's 
positions, we can broadly characterize these as falling into 
one of two schools of thought , which have been labelled as the 
'Advertising = Market Power' view and the 'Advertising = 
Information' view (Albion & Farris 1981) . These two research 
traditions differ markedly with respect to their assumptions 
about the way advertising affects consumer behavior. 

The market power perspective considers advertising a 
means of persuasion . In other words, advertising can change 
consumer tastes and artificially differentiate the product 
from close substitutes thereby insulating the firm from price 
competition. Thus, advertising enables the individual firm 



10 



and the industry as a whole to charge higher prices and 
obtain higher profits (Comanor & Wilson 1979) . Central to 
this thesis is the concept of product differentiation. 
Proponents of the monopoly view (Steiner 1978) often cite the 
example of national brands of aspirin which dominate the 
market, in spite of the fact that private-label products are 
physically identical and sell for less than half the price. 
The market power view also postulates that advertising by 
incumbents also creates barriers to entry for new firms in an ' 
industry and leads to increased concentration (Bain 1956; | 

t 

Kaldor 1950) . This model, though founded on assumptions t 
about consumer reactions to advertising, mainly focuses on 
the nature of competition within the industry. That is, the 
focus is on supplier rather than consumer behavior. 

An alternative to the change-of-tastes approach has 
developed from the economics of information theories, 
beginning with the work of Stigler (1961) . According to this 
view, advertising acts to change the information upon which 
consumer decisions are based, but not the criteria for 
judgment. For example, advertising might succeed in 
convincing a consumer that a certain product has a lower 
price, is more durable and is easier to use than another 
product, but it does not alter the consumer's beliefs about 
the relative importance of the attributes. In other words, 
it is assumed that consumers have well-defined rankings of 
desired attributes which are unaffected by advertising. 



11 



Thus, advertising simply serves to announce a product's 
existence and/or attributes. This view holds that 
advertising is a means of market competition; its function is 
to provide information to consumers, thereby increasing the 
number of known substitutes. This results in lower prices 
and reduces monopoly power (Nelson 1970,1974b; Telser 1963). 

Proponents of the information view of advertising 
distinguish between 'search' goods, whose quality can be 
evaluated prior to purchase, and 'experience' goods, whose 
quality can only be evaluated through experience after 
purchase (Nelson 1974a, 1974b, 1975) . Advertising for search 
goods provides direct information about product quality. 
There is no incentive on the part of the advertiser to 
provide inaccurate information about search goods, as 
consumers are able to recognize the difference between 
advertised and actual quality prior to purchase. Nelson has 
argued that the only information contained in advertising for 
experience goods is the fact that the brand is advertised and 
is therefore likely to provide more utility per dollar. The 
rationale for this is that producers of high quality brands 
have a greater incentive to advertise, as there is a greater 
likelihood of repeat purchases for such brands compared to 
low quality brands. Thus consumers may rationally treat the 
firms' advertising expenditures as a signal of product 
quality (Kirmani & Wright 1989; Milgrom & Roberts 1986; 
Nelson 1974) . Thus for both search and experience goods. 



12 



advertising provides consumers with valuable information 
about products and consequently reduces consumer search 
costs . 

Evidence of the price-reducing effects of advertising 
has been drawn from the prescription drug and retail eyeglass 
industries. Retail prescription drug prices were found to be 
higher in states that restricted prescription drug price 
advertising than they were in states that did not (Cady 
1976) . Price and non-price advertising restrictions were 
found to increase retail eyeglass prices by 20 to 100 percent 
(Benham 1972) . Moreover, this excess consumer cost did not 
appear to have brought about any offsetting public health or 
safety benefits. 

From the above discussion, it is clear that there are 
two general viewpoints among economists about the role of 
advertising. According to one, advertising is a means of 
persuasion and according to the other, advertising is a means 
of transmitting information. These philosophic differences 
between the economists have influenced their 
conceptualization of the relationship between advertising and 
price elasticity, which is discussed in the next section. 

The Impact of Advertising on Price Elasticity 
Theoretical Foundations 

The concept of elasticity of demand occupies a very 
important position in economics because it is a measure of 



13 



market power. The price elasticity of demand refers to the 
relative change in quantity demanded in response to a 
relative change in price. Specifically it is defined as the 
percentage change in the quantity of a product demanded in 
response to a percentage change in price. The relationship 
between advertising and price elasticity is therefore of 
central concern to economists concerned with advertising's 
effect on industry competition. 

In the market power model, advertising is postulated to 
lead to artificial product differentiation and to lower the 
perceived substitutability among competing alternatives. The 
arguments presented in these theories (Bain 1956; Comanor & 
Wilson 1979) suggest that advertising reduces the price 
elasticity of demand for firms, allowing them to charge 
prices above marginal costs and earn higher profits. On the 
other hand, the information view assumes that consumers do 
not have perfect information about product qualities, prices 
and other relevant product characteristics. In this model, 
price elasticity is taken to be a function of consumer 
awareness and qualitative knowledge about close brand 
substitutes rather than their mere existence. Advertising 
increases the number of known substitutes and also provides 
information about them. In this way, it increases price 
sensitivity and reduces monopoly power (Nelson 1970,1974a, 
1974b, 1975, 1978) . 



14 



From this discussion it is clear that the two schools of 
economic thought postulate different roles for advertising on 
the basis of certain implicit assumptions about consumer 
responses to this marketing variable. However, these 
assumptions about advertising's potential effect on consumer 
behavior remain essentially untested. In addition, past 
conceptualizations have not been able to isolate the various 
mechanisms by which advertising influences consumer price 
sensitivity. Thus it is likely that each of the two opposing 
viewpoints on advertising's potential effect on consumer 
behavior might hold under a certain set of conditions. It is 
proposed that a more general framework that takes account of 
these different relationships between advertising and 
consumer price sensitivity will be able to offer more 
insights on this issue. It is suggested in this paper that a 
more detailed analysis of consumer responses to advertising 
making use of insights from behavioral research will provide 
an improved understanding of the effects of advertising on 
the price elasticity of demand. The objective of the present 
research is therefore to study individual consumer responses 
to advertising and price changes. It is hoped that this 
integrated perspective might ultimately provide more secure 
behavioral foundations for generalizations at the firm and 
industry level, but it should be noted that this is not the 
objective of this research. Hence this research does not 



15 



attempt to answer questions regarding the implications of 
advertising for market competition versus monopoly power. 
Empirical Evidence 

There have been few empirical studies that directly test 
the relationship between advertising and price elasticity 

(Ornstein 1977) . Some researchers in economics have found 
support for the monopoly view of advertising, using data at 
the brand-level (Lambin 1976) and at the industry-level 

(Comanor & Wilson 1974) . Thus, both aggregated and 
disaggregated econometric studies have shown advertising to 
reduce the price elasticity of demand. Some marketing 
researchers who have specifically tested for the effect of 
advertising on elasticities have also found support for the 
view that advertising decreases price elasticity of demand 

(Krishnamurthi & Raj 1985) . 

However, it becomes more difficult to make general 
statements about the impact of advertising on price 
elasticity when we examine the work done by other researchers 
in marketing. Using cross-sectional and time-series data, 
Wittink (1977) found a positive relationship between 
advertising and price elasticity, consistent with the 
information approach. Other researchers who have also drawn 
similar conclusions, i.e., that advertising increases price 
sensitivity, have not measured price elasticity in response 
to advertising, but have examined price sensitivity through 
price advertising interactions (Eskin 1975; Eskin & Baron 



16 



1977; Prasad & Ring 1976) . Eskin & Baron, for example, 
report a negative interaction between advertising and price, 
i.e., the simple effect of price on quantity sold increased 
with an increase in advertising. They interpret this to mean 
greater responsiveness (higher price elasticity) to price 
when advertising is increased. However, the elasticity of 
demand depends not only on the slope but also on the original 
position on the curve. Thus it is not possible, simply by 
observing a negative interaction between price and 
advertising, to make general statements about changes in 
price elasticity in response to advertising. It is easy to 
imagine situations in which price elasticity might decrease, 
even though the analysis of variance tests show a negative 
interaction, as was observed in the Eskin & Baron study. A 
related problem stems from the disguised nature of the data 
employed by these researchers. To preserve confidentiality, 
Eskin and Baron applied a positive linear transformation to 
their sales data. Such linear transformations on the data 
will not affect the price-advertising interaction, but the 
elasticities will be affected. Due to these factors, it is 
difficult to make valid generalizations about the effect of 
advertising on price elasticity from an examination of price 
advertising interactions. 

Limita tions . Past research into the effect of 
advertising on price sensitivity suffers from certain 
limitations. The evidence appears to be mixed, but the lack 



17 



of comparability across studies on the conceptualization and 
design makes it impossible to generalize from such evidence. 
Part of the problem arises from the lack of consistency among 
authors in what they mean by 'advertising.' Variations in 
advertising content, which might have a potential effect on 
price sensitivity, have not been considered (Albion & Farris 
1980,1981; Lambin 1976). As we shall see in the following 
sections, advertising messages are not identical, and 
different types of advertising can be expected to have 
potentially different effects on price sensitivity. In light 
of this, it is meaningless to view advertising as a 
homogeneous activity and hypothesize its potential effect on 
price sensitivity without regard for the variation in 
content. The use of different types of dependent measures of 
price sensitivity with respect to advertising (e.g. price- 
advertising interactions or price elasticities) might also 
account for part of the conflicting evidence. 

Second, the causal relation between advertising and a 
measure of market power such as price elasticity is also a 
controversial point (McAulife 1987) . A correlation between 
advertising and price elasticity does not prove that 
advertising is the cause of an increase or decrease in price 
elasticity, especially because such correlational studies 
have not controlled for other possible causal factors. Even 
if a causal relationship exists, it is not clear what the 
direction of causality is, since it has also been shown that 



the optimal level of advertising depends on the price 
elasticity of demand (Dorfman & Steiner 1954) . Studies done 
in the econometric tradition, which seem to support the 
monopoly view of advertising (e.g., Comanor & Wilson 1974; 
Lambin 1976) , did not explicitly control for extraneous 
variables that might have had a potential effect on the 
results. In the absence of a controlled experiment 
specifically testing for the relationship, it is not possible 
to make any statement about its causal nature. Other 
researchers who have used controlled field experiments (Eskin 
& Baron 1977; Krishnamurthi & Raj 1985) could possibly make 
claims about causality. But in the absence of controls 
factors such as advertising content and the stage in the 
product life cycle, which could have a potential effect on 
the results obtained, they too have not been able to provide 
conclusive evidence about the causal relation between 
advertising and price elasticity. 

A Conceptual Framework 
This section presents a generalized conceptual framework 
for examining the relationship between advertising and price 
sensitivity. In this dissertation, price sensitivity is to 
be measured directly by examining price elasticities. An 
attempt is made to establish a causal link between consumer 
responses to advertising and the resulting price sensitivity; 
this is to be experimentally tested under controlled 



19 



conditions. This approach will consider the possible ways in 
which different kinds of advertising (primarily varying in 
content) affect consumer responses to advertising and the 
resulting price sensitivity. 

It is suggested that if the objective is to better 
understand the causal relationship between advertising and 
price sensitivity, an appropriate focus of study could be 
consumer choice as a function of price sensitivity. Hence 
the question, properly framed, is: how does advertising 
affect consumer choice? Various choice models and mechanisms 
have been suggested in the literature (Corstjens & Gautschi 
1983; Hutchinson 1986). One particular theoretical 
formulation of the choice process (Nedungadi 1987) is very 
insightful in this context, where the primary focus is on how 
advertising influences price sensitivity through its effect 
on consumer choice. In this model, brand choice is 
conceptualized as a two-stage process. In order to be 
selected, a brand must (a) first be included in the 
consumer's consideration set and (b) be preferred to others 
in the set. In other words, an alternative must first be 
brought to mind and actively considered and must also be 
perceived to offer the maximum utility among all members of 
the set. It follows that the probability of a brand being 
finally chosen can be changed by influencing its inclusion in 
the consumer's consideration set or by altering the perceived 
utility of that brand relative to others in the set. 



20 



Lynch and Bloom (1987) use this conceptualization of 
choice in their proposed framework for examining the effects 
of advertising on consumer price sensitivity. They suggest 
that advertising could influence the probability of choice by 
determining the size of the consumer's consideration set . 
Advertising that increases the probability of inclusion of 
the advertised brand in the consideration set increases the 
size of the set, and advertising that inhibits the 
probability of simultaneous inclusion of competing brands, 
reduces the size of the consideration set. In short, 
advertising could have either a positive or negative effect 
on the size of the consideration set. All else equal, the 
elasticity of demand for any given brand to changes in its 
price will be greater the larger the number of brands that 
are simultaneously considered. 

Advertising could also potentially influence the 
dispersion of brand-utilities . It could increase the 
perceived utility for the sponsored brand and decrease the 
perceived utilities for competing brands. Advertising could 
change perceptions about market offerings or alter the 
criteria consumers use to evaluate products . In this way, by 
changing the consumer's preferences for the various 
alternatives, advertising could have a positive or negative 
effect on the probability of choosing the advertised brand. 



21 
Advertising-induced Changes in the Size of the Consideration 

The consideration set consists of the group of brands 
that the buyer actively considers when making a choice 
(Campbell 1969) . These brands become the alternatives in the 
buyer's choice decision. It has been suggested (Alba and 
Chattyopadhay 1985; Hauser & Wernerfelt 1990; Nedungadi 1987) 
that the consumers faced with an assortment of brands attempt 
to concentrate on a subset of alternatives in order to gain 
efficiency in shopping. Therefore, marketers need to 
organize their efforts in such a way that their particular 
brand is included in the consumer's consideration set. The 
consideration set is a dynamic entity varying over different 
types of choice situations (Nedungadi 1987). Thus, the 
marketer's task involves more than simply ensuring consumer 
awareness of the brand's existence; the brand has to be 
retrieved and considered at the time of choice. Advertising 
could ensure that the advertised brand is included and also 
prevent simultaneous inclusion of other brands. Thus 
advertising could have a positive or negative effect on the 
size of the consumer's consideration set. 

In the real world, there are many instances of an 
advertiser attempting to place the sponsored brand in the 
consumer's consideration set. For example, reminder 
advertising often attempts to provide retrieval cues at the 
time of choice (Keller 1987) . Such efforts might ensure that 
the advertised brand is included in the consumer's 



22 



consideration set. Empirical research has also shown that 
advertising for a brand can substantially inhibit the recall 
of competing brands (Alba & Chattyopadhyay 1986) . This has 
important implications for situations in which an alternative 
has to be remembered to be considered. For example, when 
consumers decide which stores they want to patronize, all the 
alternatives are not arrayed in front of them. Thus, they 
have to retrieve relevant information from memory. In this 
case, if advertising for a few stores dominates the media, 
these stores are likely to be salient in the consumer's mind 
relative to unadvertised stores. In this way, advertising 
might ensure inclusion of the advertised brand for 
consideration and thus influence the choice decision by 
altering the size of the consideration set (Silk & Urban 
1978) . 

The effects of advertising on consumer decisions allow 
us to derive its implications for revealed price sensitivity. 
If price is considered to be an argument in a multi-attribute 
utility function (Huber et al . 1986), then all other things 
being equal, the greater the number of brands in their final 
consideration set, the more price sensitive consumers are 
likely to be (Lynch & Bloom 1987) . Thus, if advertising 
increases the size of the consideration set by ensuring 
inclusion of advertised brands, it leads to increased price 
sensitivity. This implies that advertising for a particular 
brand leads to increased price sensitivity for competing 



23 



brands that are already included in the consideration set. 
Advertising for a brand might also inhibit the inclusion of 
potential competitors in the consideration set and thereby 
result in reduced sensitivity to changes in its own price. 
These changes in consumer price sensitivity caused by 
advertising are mediated by the size of the consideration 
set. In this study, all brands are advertised and therefore 
the institution of advertising results in an increase in the 
size of the consideration set. 

From the preceding discussion it is clear that the 
consideration set is an important construct for examining the 
effect of advertising on consumer choice. One possible way 
in which advertising could influence consumer price 
sensitivity for the advertised brand is simply by changing 
the size of the consideration set in which this brand is 
included. The effects just described are purely the result 
of memory factors. The next section discusses changes in 
price elasticity that result from advertising- induced changes 
in preference. 

Advertising-i nduced Changes in Preference given Inclusion in 
the Consideration Set 

Economic analysis generally treats advertising as a 

homogeneous activity that is evaluated independently of why 

it might increase demand (Leffler 1982) . Yet, the effect of 

advertising need not be the same in different markets or in 

different situations within a market. For example, price 



24 



comparison ads of standardized products might increase price 
sensitivity both by providing more information as suggested 
by economic analysis and by the behavioral mechanism of 
increasing salience of price and inhibiting the salience of 
non-price attributes (Bettman & Sujan 1987; Feldman & Lynch 
1988; Lynch & Bloom 1987). 'Image' advertising of 
heterogeneous, differentiated products might reduce price 
sensitivity by increasing the importance of non-price 
attributes in the consumer's utility function. Hence, an 
examination of this relationship should be prefaced by the 
particulars of the products advertised and the message 
delivered by the ad. If advertising is a multif aceted, 
heterogeneous activity, general statements as to the economic 
effects of advertising might not be possible or 
interpretable . Rather, it would be more reasonable to 
examine systematically conditions under which advertising can 
be expected to have a positive or negative effect on price 
sensitivity . 

In the economic literature (and even in the marketing 
literature which has attempted to examine price-advertising 
interactions) , the content of advertising has never been 
considered. Rather, the amount of advertising has been 
emphasized. However, perhaps the content of advertising is 
what is important in terms of its potential effect on price 
sensitivity, and this is an important factor to be considered 
in the examination of this relationship (Albion & Farris 



25 



1981; Lambin 1976) . It may be meaningless to judge the 
information potential of advertising on consumer behavior 
without regard to its content. In this regard, it would be 
too simplistic to postulate, as some economists (Nelson 1970, 
1974, 1978) have done, that all advertising is informative 
rather than manipulative or persuasive, because this assumes 
that consumers have fixed preferences and the only objective 
of the ad is to provide functional information. Based on 
evidence from the behavioral literature, some authors suggest 
that an ad for a brand can change perceptions of the product 
and hence can affect the probability of choice (Alba & 
Hutchinson 1990; Bettman & Sujan 1987; Hoch & Ha 1986; Lynch 
& Bloom 1987) . 

As noted above, the probability that a consumer chooses 
a particular brand depends upon his/her preference or utility 
for the brand relative to competing alternatives (Cortsjens & 
Gautschi 1983; Nedungadi 1987) . Research on consumer 
preferences and attitudes has relied heavily on the use of 
multi-attribute models. Within the general paradigm of 
multi-attribute models, one can envision several possible 
ways in which an attitude toward a particular brand can be 
changed (Lutz 1975; Wilkie & Pessemier 1973) . Basically, 
these models postulate that individuals' attitudes toward a 
brand is some function of their beliefs about the extent to 
which the brand possesses certain attributes, and their 
evaluation of the importance of these attributes. Thus, 



26 



attitudes are construed as a linear combination of brand 
beliefs and importance-weights on certain attributes. 

Marketers often attempt to use advertising to increase 
the perceived utility for the sponsored brand. One of the 
key benefits of the multi-attribute approach is that it 
offers useful guidance in this regard. Consumer researchers 
suggest several advertising strategies which can improve 
consumer attitudes toward the advertised brand relative to 
competitors (Boyd et al . 1972; Lutz 1975). In line with the 
theoretical formulation stated above, all these strategies 
attempt to improve consumer preferences for the advertised 
brand in one of the following ways: (a) by changing existing 
beliefs about the brand or competing brands, (b) by 
increasing the importance of an attribute on which the brand 
is strong, (c) by decreasing the importance of a weak 
attribute, and (d) by adding an entirely new attribute (Boyd 
et al. 1972) . 

In the preceding paragraphs, the various possible ways 
in which advertising might change consumer preferences for 
the sponsored brand have been outlined. This leads to an 
increased likelihood of it being chosen, if it is already 
included in the consideration set. This is likely to have 
implications for price sensitivity. The mechanisms by which 
they could occur are discussed below. 

First, advertising for the sponsored brand could change 
the distribution of preferences and create a situation in 



27 



which the ordering of preferences is not sensitive to changes 
in price. The greater the dispersion in the consumers' mind 
among the utilities of the alternative brands, the less price 
elastic should be their demand for any individual brand. 
Advertising might lead to some brands being perceived to be 
clearly superior relative to others. In other words, 
advertising, by increasing the perceived utility for the 
advertised brand relative to that of competing brands, might 
lead to a lower price elasticity for the sponsored brand. 

A second way in which advertising might potentially 
reduce price sensitivity is by increasing the salience of 
non-price attributes and thereby decreasing the importance of 
price. Thus, advertising can cause consumers to 'frame' 
their choices in terms of the evaluative criteria suggested 
by advertisers (Alba & Hutchinson 1990; Bettman & Sujan 1987; 
Hoch & Deighton 1989) . It has been reported that increasing 
the salience of product attributes suppresses the ability to 
recall unmentioned attributes (Alba & Chattyopadhyay 1985) . 
Also, by increasing the salience of a particular attribute, 
advertising might ensure that this attribute is used in 
subsequent evaluations, and thereby increase the effective 
weight given to that attribute (Feldman & Lynch 1988) . Taken 
together, this suggests that advertising, by increasing the 
salience of non-price attributes, could lead to decreased 
price sensitivity. 



28 



The preceding paragraphs have discussed the role of 

advertising on consumer price sensitivity mediated by altered 

preferences as a result of advertising. It has been pointed 

out that advertising could lead to decreased price 

sensitivity for the advertised brand either by enhancing the 

importance of non-price attributes in consumer choice or by 

changing the perceived overall utility for that brand and the 

difference in utility between that brand and the consumer's 

most preferred brand. Some authors suggest that this would 

be less true for consumers with higher levels of expertise 

(Alba & Hutchinson 1990; Lynch & Bloom 1987) 

Interplay between Adve rtising-induced Preference and the 
Consideration Set 

The framework so far has conceptualized the impact of 
advertising on price sensitivity as being mediated by changes 
in either the size of the consumer's consideration set or the 
perceived utilities of various brands. From this perspective 
of the choice process and the role advertising plays in it, 
it might seem as if the size of the consideration set is 
independent of consumer preferences for the competing 
alternatives. However, the perceived utilities of the 
different product offerings, aside from having a direct 
effect on price sensitivity, might also limit the size of the 
consumer's consideration set. 

Through the direct effect on preferences, advertising 
might also have an indirect effect on the size of the 



29 



consideration set. Ad-induced changes in utility may cause 
some alternatives to become dominated and, therefore, removed 
from the consideration set. In other words, advertising, by 
changing brand preferences, might determine the number and 
type of brands that the consumer considers at the time of 
choice. Advertisers often attempt to differentiate the 
product on certain advertised dimensions. The increased 
salience of this differentiating attribute might lead to a 
greater attention being given to that attribute at the time 
of choice (Gardner 1983) . The prominence given to the 
attribute might have an effect on the choice heuristic 
employed by consumers. If noncompensatory choice processes 
are modelled as though it were compensatory, this would 
affect the revealed importance of the attribute (Johnson & 
Meyer 1984) . 

The market power view of advertising asserts that 
advertising differentiates products and creates a situation 
in which the advertised brand is perceived to have few close 
substitutes . This might be taken to imply that when the 
advertised brand is included in a consumer's consideration 
set, few other brands are simultaneously included. This is 
because the perceived benefit from including additional 
brands in the consideration set will be less when some 
clearly preferred brand is already included (Hauser & 
Wernerfelt 1990) . Since price sensitivity is a function of 
the perceived substitutability among brands, differentiating 



30 



advertising, by changing preference structure, reduces the 
size of the consideration set. This leads to decreased price 
elasticity. In this case, it might also affect the 
composition of the set, such that the variance in utilities 
for brands in the consideration set is less than the variance 
for all brands in general. This might imply higher cross- 
price sensitivity for brands included in the consideration 
set . 

In this section, the interdependence between preference 
and the size of the consideration set was discussed. 
Advertising, by changing the relative utilities for the 
various product offerings, is likely to alter the size and 
composition of the consideration set, which in turn has 
implications for price sensitivity. 
Summary 

Prior conceptualizations regarding the relationship 
between advertising and consumer price sensitivity have been 
too simplistic in that they have looked for main effects of 
advertising. Thus the literature remains confusing and the 
disputes unresolved. It is proposed that we can arrive at a 
better understanding of the phenomenon by looking at higher- 
order interactions involving many other factors and a more 
sophisticated treatment of consumer choice behavior. 

The proposed framework discussed above attempts to 
isolate various possible causal mechanisms by which 
advertising affects consumer price sensitivity by affecting 



31 



the awareness of substitutes and the ability to remember 
substitutes of which one is aware. First, it might change 
the size of the consideration set by affecting the awareness 
of substitutes and the ability to remember substitutes of 
which one is aware. Second, it might alter preferences for 
the advertised brand given inclusion in the consideration 
set. Third, it might also reduce the size of the 
consideration set through preferential mechanisms. The 
greater the variance in utilities, the less the incentive to 
consider brands other than one's most preferred ones, in the 
hope that a price discount for a less-liked brand would cause 
it to be preferred to the normally favored brands. 

The next chapter details an experiment that is designed 
to empirically test the concepts discussed in the framework. 



CHAPTER III 

AN EMPIRICAL INVESTIGATION OF THE IMPACT OF ADVERTISING 
ON PRICE ELASTICITY: CONCEPTUAL HYPOTHESES AND STUDY DESIGN 



Objective 
This study attempted to examine the relationship between 
advertising and price sensitivity mediated by two key 
constructs: the size of the consideration set and the 
dispersion of brand-utilities. 

Background 
The Consideration Set 

Consumers are faced with a multiplicity of alternatives 
in course of making their regular purchase decisions. In 
each product category, consumers are aware of a large number 
of brands and have to make their choice from among these 
brands. To simplify their decision making, they must make 
their selection from a smaller group of brands (Alba & 
Chattyopadhyay 1985; Hauser & Wernerfelt 1990; Nedugadi 1987; 
). In other words, the consumer's final choice is made from 
the consideration set — a certain subset of alternatives from 
the total number of options available in the market. This 
definition is similar to the 'evoked set' notion used by 
marketing researchers (Campbell 1969; Howard & Sheth 1969; 
Narayana & Markin 1975; Parkinson & Reilly 1979) . 



32 



33 



The concept of a consideration set is important for the 
understanding of consumer choice. First, it emphasizes that 
mere awareness of a brand is not sufficient for it to be a 
candidate in the consumer's final choice set. The brand has 
to be actively considered at the time of decision making. 
The brands that consumers consider at the time of choice form 
a small subset of the total number of alternatives of which 
they are aware (Silk & Urban 1978) . Second, the 
consideration set is not a static entity. Of course, 
research in consumer behavior has often implicitly assumed 
that the consumer chooses from a fixed set of brands. 
Typically, consumers are presented with a set of brands and 
information about them, and are asked to make a choice or to 
indicate their preferences (Lynch & Srull 1982) . But in 
practice, the size and composition of the consumer's 
consideration set is determined by certain voluntary and 
involuntary factors. The size of the consideration set is a 
positive function of the consumer's ability and motivation to 
retrieve and consider a certain number of brands for 
evaluation at the time of choice. Thus anything that 
positively affects the ability of the consumer to retrieve 
brands from memory will increase the size of the 
consideration set, and factors that negatively affect this 
ability will tend to reduce the size of the consideration 
set. Similarly, factors that alter the motivation of the 
consumer to retrieve information from memory will have an 



34 



effect on the size of the consideration set. For example, 
the inclusion of highly preferred brands in the consideration 
set is likely to reduce the motivation to consider additional 
brands for evaluation and choice. This will reduce the size 
of the consideration set and also affect its composition. 
Hence the consideration set is likely to vary over time, 
depending on whether factors that are present when the choice 
is made result in an increase or decrease in the ability and 
motivation to retrieve brands for consideration. 
Presence or Absence of Advertising 

Advertising for the brand is an important factor that 
can influence brand choice by affecting the ability and the 
motivation to place the sponsored brand in the consideration 
set. In today's competitive environment, the consumer is 
faced with an assortment of brands and has to focus on a 
subset of brands in order to gain efficiency in shopping. 
Under such a scenario, marketer-controlled advertising can 
serve important purposes that help consumers make the choice 
decision from among the numerous alternatives that they 
encounter . 

This study will consider the effects of the institution 
of advertising on price elasticity. In other words, this 
research will compare the case when none of the brands 
advertise to that when all brands engage in advertising. 

Reminder advertising . First, advertising can result in 
keeping the sponsored brand salient in the consumer's mind. 



35 



The ability to recall a brand depends on the accessibility of 
the brand in memory. It has been noted earlier that mere 
awareness of the brand is not enough for it to be considered 
at the time of choice. In order to be chosen, the brand must 
be retrieved from long-term memory and included among the 
subset of alternatives that comprise the consumer's 
consideration set. Advertising for even well-known brands 
keeps the brand salient in the consumers mind and increases 
the probability that it would be retrieved at the time of 
choice (Hauser and Wernerfelt 1990; Silk & Urban 1978;). 

Consider a situation in which the marketers use 
advertising for their brands, but the advertising is of a 
reminder nature. That is, its major objective is to keep the 
brand name accessible in memory, and thereby increase the 
probability that the sponsored brand is included in the 
consumer's consideration set. In fact, some authors note 
that in many instances, the aim of advertising is to place 
the brand in the consumer's evoked set (Wilkie & Farris 1976, 
Wittink 1977) . Though the evoked set they are referring to 
is more akin to an 'awareness set', it is possibly true that 
keeping the brand salient in the consumer's mind through 
reminder advertising also enables it to be considered at the 
time of choice. In other words, reminder advertising 
provides retrieval cues at the time of choice, and thereby 
increases the size of the consideration set. 



36 



As stated in the earlier section, this study will 
consider the effects of advertising when all brands are 
advertised. If all firms pursue this strategy, there should 
not be any inhibition effects resulting from the salience of 
a few brands in memory as has been observed by some 
researchers (Alba & Chattyopadhyay 1986) . In such a case, 
reminder advertising of all brands will increase the salience 
of brands in memory and thereby increase the size of the 
consideration set. 

Differentiatin g advert! si ng . Marketers might also try 
to use advertising to differentiate their brands from 
competitive offerings and thereby reduce the number of 
perceived substitutes. In the real-world, advertising is 
often used to position a parity product away from substitutes 
with the objective of insulating the brand from competition. 
This type of advertising leads to an increase in the 
dispersion of brand-utilities among the alternatives in the 
consideration set. In other words, differentiating 
advertising will result in some brands being perceived to be 
clearly superior to others and thereby increase the 
probability of these brands being chosen (Hauser & Wernerfelt 
1990) . 

If the focus was on the variations in a single brand's 
advertising, this could either increase or decrease 
interbrand variance of interbrand utilities. For example, if 
the top brand advertises, then this increases variance of 



37 



utilities. If a less attractive brand advertises, this will 
decrease the variance. In this study, all brands advertise. 
Under such a scenario, information contained in 
differentiating advertising will result in an increase in the 
dispersion of utilities compared to the situation when none 
of the brands advertise. 

Besides affecting preferences for the brands in the 
consideration set, differentiating advertising can affect the 
size of the consideration set itself. As noted earlier, 
consideration set size is a function of not only the ability, 
but also the motivation of the consumer to retrieve brands 
from memory. Differentiating advertising might be expected 
to positively affect the ability to retrieve brands by 
keeping brand names salient in memory. Thus, compared to the 
situation in which none of the brands are advertised, 
differentiating advertising, by increasing the ability to 
recall brand names, is expected to increase the size of the 
consideration set. 

However, by directly affecting the distribution of 
perceived utilities for the various brands, differentiating 
advertising is also likely to have an indirect negative 
effect on the motivation to retrieve brands for 
consideration. The higher the utility from the best brand in 
the evoked set, the less likely is the consumer to consider a 
new brand (Hauser 1989) . Thus, by increasing the utilities 
and the variation in utilities of brands in the consideration 



38 



set, differentiating advertising is likely to reduce the 
consumer's motivation to retrieve brands for consideration at 
the time of choice. 
The Decision Environment 

The environment in which the choice is made has a 
potential role in moderating the effects of advertising on 
the consideration set and the dispersion of brand-utilities. 
Decision environments differ in terms of the degree to which 
the consideration set is driven by external factors such as 
salesperson's recommendations or point-of-purchase display, 
or by the information that the consumer has to access from 
memory. 

Consider the case of the consumer who is deciding on the 
store at which to shop. Decisions such as this are 
necessarily memory-based. Consumers do not have all the 
alternatives available to them at the time of choice, and 
therefore must retrieve the information from memory. This is 
the pure 'memory-based' choice situation, in which the 
environment does not provide any of the relevant information 
(on brand names or attributes) at the time of choice, and the 
consumer has to retrieve the information from memory. 

There are other situations in which the consumer could 
rely on external cues (in whole or in part) for the 
generation of alternatives. For example, consider the 
consumer walking down a supermarket aisle. In this 
situation, the consumer has brand and attribute information 



39 



on all the alternatives readily available at the time of 
choice. In such a case, the environment is providing all the 
information relevant to making the decision. This is what we 
term 'stimulus-based' choice environment. In such a 
situation, the consideration set is likely to be determined 
by external cues provided by the environment, e.g., price 
rebates or end-aisle display. 

Even in the case of stimulus-based choices, as in our 
supermarket example earlier, it may be argued that the 
consumers do not spend their time exhaustively considering 
all possible alternatives. Even in such situations which are 
in principle, purely stimulus-based, it is contended that 
memory factors play a crucial role (Alba et al. 1990) . 
Research in consumer behavior (Hoyer 1984; Park et al 1989) 
suggests that consumers do not engage in elaborate processing 
of package information at the time of choice. Thus it is 
likely that consumers retrieve relevant information from 
memory in order to concentrate on a subset of alternatives 
and thus gain efficiency in shopping. 

In short, for certain decisions, consumers might have to 
retrieve information from memory in order to evaluate and 
choose from a candidate set of brands. In other situations, 
they could rely on external cues to some extent. The more 
the consumer relies on external cues to guide choice, the 
more the decision is said to be stimulus-based. But, as 



40 



noted earlier, the pure stimulus-based choice decision is a 
rare phenomenon . 

In this context, it would be useful to distinguish 
between memory-based versus stimulus-based specification of 
attributes and brands. The few studies (Alba & Marmorstein 
1987; Biehal & Chakravarti 1983,1986; Lynch et al . 1988) 
which have considered memory-based decision making have 
assumed that brand names are given and that relevant 
information on attributes must be accessed from memory. This 
is representative of situations in which consumers are 
shopping from store to store, where not all the attribute 
information is available in front of them, and they have to 
retrieve relevant information about the brands from memory in 
order to make a judgment or choice. But there are other 
situations in which even the brand name has to be generated 
from memory, irrespective of whether the attribute 
information is readily available or not. As noted above, the 
consumer might be faced with a variety of choice environments 
which offer different types of information at the time of 
decision making. 
Implications 

Whether the choice is stimulus-based or memory-based has 
implications for the alternatives considered and the 
informational inputs that are used in decision making 
(Nedungadi 1987) . When the choice is memory-based, the size 
of the consideration set will depend on the consumers' 



41 



ability and motivation to retrieve brands from memory. If 
there is no advertising for the various brands, then the 
consideration set in a memory-based choice situation depends 
purely on the consumer's capacity to retrieve brand 
alternatives. Hence the size of the consideration set is 
likely to be smaller in the memory-based environment than in 
the stimulus-based environment. 

When the decision is such that brand names have to be 
remembered to be considered, the consumer has to rely on 
memory to retrieve brands for consideration. Therefore 
marketers try to organize their efforts in such a way that 
their particular brand is considered at the time of choice 
(Narayana & Markin 1975) . Thus, under memory-based 
situations, reminder advertising will lead to an increase in 
the ability to recall brands and thereby increase the number 
of brands considered for choice. If all firms pursue this 
strategy, there should not be any inhibition effects 
resulting from the salience of a few brands in memory as has 
been observed by some researchers (Alba & Chattyopadhyay 
1986) . 

Differentiating advertising has a positive effect on the 
ability to recall brands and can therefore increase the size 
of the consideration set. But this type of advertising is 
also likely to negatively affect the motivation to recall 
brands for consideration. In other words, if this type of 
advertising results in some brands being perceived as clearly 



42 



superior to others, it is unlikely that consumers will 
attempt to include clearly dominated alternatives in their 
consideration sets. Research in consumer behavior has shown 
that more preferred brands are likely to be more easily 
retrieved, simply because preference leads to brand usage and 
the rehearsal that this produces increases the brand's 
accessibility in memory (Nedungadi & Hutchinson 1985) . Thus 
consumers are more likely to retrieve brands that they 
perceive will provide the maximum utility (Hauser & 
Wernerfelt 1990) . Therefore, the size of the consideration 
set is likely to be negatively related to the perceived 
superiority of the brands retrieved initially. Hence, in 
memory-based choice situations, differentiating advertising, 
if it leads to changes in brand perceptions and preferences, 
might lead to a reduction in the size of the consideration 
set . 

When choice is stimulus-based, the environment provides 
the relevant information, and reminder advertising is not as 
important a determinant of the size of the consideration set. 
This is because the reminder function of advertising is not 
expected to be crucial in helping the consumer decide on the 
brands to consider before making a choice. However, it 
should be recognized that many real world choice environments 
in which stimulus information is available, consumers ignore 
it and make decisions based on memory. 



43 



Differentiating advertising will lead to an increase in 
the variance of interbrand preferences. This is likely to 
lead to a reduction in the size of the consideration set if 
consumers concentrate on the highly preferred alternatives in 
order to gain efficiency in shopping. 

Advertisers use differentiating advertising to alter 
preferences for the various brands. Consumers might use the 
criteria provided in the ad for the purpose of evaluation and 
choice (Bettman & Sujan 1987; Gardner 1983, Hoch & Deighton 
1989) . Such 'framing' effects are likely to limit the size 
of the consideration set. Thus, in stimulus-based 
situations, a differentiating advertising strategy might lead 
to greater attention being devoted to certain attributes, 
thereby changing the composition of the consideration set. 
For example, the consumer might use a particular attribute in 
order to eliminate alternatives, and the members of the final 
consideration set might be similar on that particular 
attribute . 

The greater the number of brands in the consideration 
set, the greater is the price sensitivity, because the 
existence of a greater number of substitutes decreases the 
utility gap between the most preferred brand and its 
competitors thereby increasing the elasticity of demand to 
price cuts by the latter (Lynch & Bloom 1987) . The positive 
effect of advertising on price sensitivity will be larger if 
consideration set sizes in the absence of advertising were 



44 



smaller. That is, if advertising increases consideration set 
sizes from 2 to 3, this will have a more dramatic effect on 
price elasticity than if it increases consideration set sizes 
from 4 to 6 . Thus, under memory-based choice situations, 
reminder advertising, which increases the size of the 
consideration set, tends to increase price sensitivity, and 
differentiating advertising, which reduces the size of the 
consideration set, tends to decrease price sensitivity. 
Differentiating advertising for brands will also increase 
relative preference for some and thus have a negative effect 
on price sensitivity. 

Even when advertising has no effect on the size of the 
consideration set, it can lead to an increase in the 
dispersion of brand-utilities. Advertising that seeks to 
convey attribute information and thereby differentiate the 
sponsored brand from close substitutes often results in 
changing the relative overall preference for the various 
brands. The larger the variance in utilities, the larger is 
the price discount necessary to change purchase patterns. 
Therefore differentiating advertising would be expected to 
increase the dispersion of brand-utilities and thus lower the 
price elasticity in the stimulus-based choice environment. 
Assumptions 

1 Holding constant the size of the consideration set, 
price elasticity is inversely related to the dispersion 
of brand-utilities. 



45 



2. Price elasticity is a positive function of consideration 
set size. 

3a. Differentiating advertising conveys more utility- 
relevant information than reminder advertising. Thus, 
it can be expected to have greater effects on interbrand 
variance in utilities than reminder advertising. 

3b. Reminder advertising will have minimal effects on the 
interbrand dispersion of utilities compared to a 
situation of no advertising. 

4. The size of the consideration set is a function of 
certain voluntary and involuntary factors. It is 
positively related to the ability and motivation of the 
consumer to retrieve a certain number of brands from 
memory. Anything that increases (decreases) the ability 
or the motivation to retrieve brands for consideration 
will increase (decrease) the size of the consideration 
set . 

5. Both reminder and differentiating advertising can 
facilitate retrieval of brands. 

6. When the environment provides the names of all relevant 
brands (stimulus-based environment) , the primary effect 
of advertising on the consideration set will be through 
its effect on the dispersion of preferences. 

Specifically, 

6a. Reminder advertising of all brands will not have a 
significant effect on the size of the consideration set 



46 



to the extent that it will have minimal effects on the 
dispersion of brand-utilities. 
6b. Advertising that seeks to differentiate the brands 
serves to change brand-utilities and the inter-brand 
variance of utilities, and thereby reduce the size of 
the consideration set . 

7. When the consumer must remember the names of all 
relevant brands (memory-based environment) , there will 
be a large simple effect of advertising on the size of 
the consideration set . 

Specifically, 

7a. Consideration set sizes will be larger when advertising 

is of a reminder nature than if no brands advertise. 
7b. Consideration set sizes will be smaller when all brands 

use differentiating advertising than if no brands 

advertise . 

8. Average size of the consideration set will be larger 
when choice is stimulus-based than when choice is 
memory-based . 

Hypotheses 

HI. When choice is stimulus-based, price sensitivity will be 
greater than when choice is memory based. 

H2 . When choice is stimulus-based, the only effect of 
advertising on price sensitivity is through a change in 
perceived utilities of the various brands. 



47 



H2a. When choice is stimulus-based, differentiating 
advertising will result in decreased price sensitivity 
compared to a situation of no advertising. 

H2b. When choice is stimulus-based, there will be no 
significant effect of reminder advertising on price 
sensitivity compared to a situation of no advertising. 

H3 . When choice is memory-based, there will be a significant 
effect of advertising on price sensitivity. 

H3a. When choice is memory-based, advertising that seeks to 
differentiate the brand from competing brands will 
result in decreased price sensitivity compared to a 
situation of no advertising. 

H3b. When choice is memory-based, reminder advertising of all 
brands will result in greater price sensitivity compared 
to a situation of no advertising. 

Overview of Research Methodology 
Three experiments were designed to test the hypotheses 
concerning advertising and price elasticity, mediated by the 
two constructs of interest, namely, size of the consideration 
set and the dispersion in preferences for the various brands. 
The first two experiments served as pretests for Experiment 3 
which examined the effect of advertising on price 
elasticities. In order to help the reader understand the 
issues in the design of Experiment 3 that Experiments 1 and 2 



48 



were intended to resolve, a brief overview of Experiment 3 is 

provided first . 

Design 

This section presents the overall methodology employed 
in the Experiment 3, which utilized a computer-based shopping 
simulation. Three factors were manipulated in the 
experiment. A 2X3X12 design was used, with Decision 
Environment (Stimulus-based / Memory-based) and Advertising 
Condition (Differentiating / Reminder / No Ads) as between- 
subjects factors. Prices of twelve brands were manipulated 
(High / Low) within-sub jects , independently of each other, 
allowing independent estimation of the price elasticities of 
the 12 brands. These price elasticities served as the 
primary dependent measures. Subjects participated in groups 
of one to four, with each group being randomly assigned to 
one of the six cells of a 2X3 between-sub jects design. 
Procedure 

Subjects were initially exposed to attribute information 
on twelve brands of Canadian candy bars that were previously 
unfamiliar to them. After demonstrating by a recognition 
task that they were 'aware' of the brands, they were 
dismissed. They came back to the laboratory at a later point 
in time, when they were required to go on 16 shopping trips 
for candy bars . 

The subjects' objective on each of the 16 shopping trips 
was to maximize satisfaction given a budget of $3.00 per 



49 



trip. Their task was to allocate the amount among candy bars 
of their choice. They were told that the prices of the 
various brands would change from one shopping trip to 
another. Subjects were also told that they would actually 
receive the candy bars purchased on one of these sixteen 
shopping trips. These aspects of the design and procedure 
were the same for all subjects. 
Experimental Factors 

Advertising condition. The type of advertising was 

manipulated at three levels. Subjects in the differentiating 
advertising condition saw ads for all twelve candy bars. 
These ads provided information on brand attributes and were 
predicted to change subjects' relative preference for the 
brands compared to a situation where none of the brands were 
advertised. Subjects in the reminder ad condition saw ads 
that provided information on the package and brand name, 
similar to some print or billboard advertising. These ads 
were predicted to increase the likelihood of inclusion of the 
advertised brand in the consideration set but to have minimal 
effect on preferences. Subjects in the no advertising 
condition served as a control group, and were not exposed to 
any ads for the candy bars prior to going on the 16 shopping 
trips . 

D ecision environment . The decision environment was 

manipulated at two levels. Subjects in the stimulus-based 
choice condition were provided with a list of all twelve 



50 



brands to use during shopping trips. They could therefore 
use this list to decide which brands to examine for price 
information and purchase. This was intended to reflect real- 
world situations in which the aisle display or a salesperson 
reminds the consumer of the set of brands that might actually 
be purchased. 

Subjects in the memory-based choice condition did not 
have such a list available to them. These subjects could 
only search for price information on brands that they were 
able to recall on their own. Therefore they could only 
purchase brands that they could retrieve from memory. This 
manipulation was intended to reflect real-world decision 
situations in which the consideration set is typically 
generated from memory. 

Chapter IV describes the pretests (Experiments 1 and 2) 
that tested some assumptions arising from the above 
discussion. Chapter V details the methodology employed in 
Experiment 3 to test the hypotheses on the impact of 
advertising on price sensitivity. Chapter VI presents a 
discussion of the results of Experiment 3. Chapter VII 
develops the implications of these findings for research in 
consumer behavior. 



CHAPTER IV 

PRETESTS 

Overview 

The hypotheses developed in the last chapter concerned 

how two basic theoretical constructs mediate the effects of 

advertising on price elasticity. These two mediating 

variables are the size of the consideration set and the 

dispersion of brand-utilities. It was argued that 

advertising could lead to an increase or decrease in the size 

of the consideration set, with resultant effects on price 

elasticity. Also, advertising that seeks to differentiate 

brands will increase the variance in perceived utilities for 

the various brands and thereby lower price elasticity. 

This chapter discusses two pretests which were designed 
to examine the effect of advertising on the ability and 
motivation to consider brands for evaluation and choice. 
These experiments were expected to provide insights on the 
hypothesized mediating roles of the theoretical constructs 
discussed in the framework. The following specific issues 
were investigated: 

1. the effects of advertising on the number of brands 
recalled; 

2. the effects of advertising on the number of attributes 
recalled; 



51 



52 



•3. the effects of advertising on interbrand differences in 
preference; 

4. the effects of advertising on price perceptions; 

5. the effects of advertising on interbrand differences in 
perceptions . 

Experiment 1 
This sections details the above mentioned issues and the 
predictions that were tested in the first experiment. The 
subjects, the methodology and the stimuli are discussed. The 
results of the pretest are also presented. 
Key Issues Investigated 

The issues that were investigated in these pretests are 
presented below. 

Effects of adve rtising on the number of brands recalled . 
It has been suggested in this paper that one of the major 
roles of advertising is to remind, i.e., to provide recall 
cues in order to ensure that the brand is retrieved for 
evaluation. The reminder function of advertising becomes 
especially important in situations in which the brand has to 
be remembered to be considered. Advertising is expected to 
have a significant effect on the number of brand names 
recalled in the memory-based decision environment. In the 
absence of advertising, subjects have to rely on memory to 
recall brand names for consideration. Advertising (both 
reminder and differentiating) can therefore be expected to 



53 



increase the ability of the consumer to recall brand names by 
making them more accessible. Both differentiating and 
reminder advertising are likely to have similar effects on 
free recall of brand names. On the basis of the above 
discussion the following prediction was made: 
PI. There will be a significant effect of advertising on the 

number of brand names recalled correctly. 
Thus the proposition tests for significant differences in 
memory for brand names when subjects are exposed to 
advertising compared to the situation in which none of the 
brands are advertised. 

In the last chapter it was hypothesized that under 
memory-based decision environments, reminder advertising will 
increase the size of the consideration set. For such an 
effect to hold, it was important to ensure that even though 
subjects were aware of the brands, the names were not readily 
accessible in memory. If the above proposition is supported, 
it implies that advertising could be expected to enhance the 
ability to recall and thereby influence the size of the 
consideration set. 

Effects of advertising on the number of attributes 

recalled. It is often suggested that advertising is a source 
f valuable information on brand attributes. If this is 
true, then differentiating advertising, besides providing 
cues for brand name recall, will also result in making 
attribute information more salient in memory. In the absence 



o 



54 



of advertising, subjects will have to rely on the initial 
(product and package) information on the various brands in 
order to retrieve attribute information from memory. 
Differentiating advertising, if it provides brand attribute 
information should increase the ability to recall brand 
attribute information and thereby have a positive effect on 
the number of attributes recalled correctly. Therefore the 
following predictions were made: 
P2a. Differentiating advertising will result in a greater 

number of attributes recalled compared to a situation of 

no advertising. 
P2b. There will be no significant effect of reminder 

advertising on the number of attributes recalled 

correctly . 
Thus this prediction tests for differences in the salience of 
brand attribute information across the three advertising 
conditions. It was expected that such information was likely 
to form the basis of perceived utility differences across the 
brands . 

It was postulated in this paper that differentiating 
advertising will lead to an increase in the dispersion of 
brand-utilities. As stated earlier, perceived utilities for 
various brands are likely to be a function of the information 
that is available at the time of choice. Subjects in this 
experiment had access to two sources of brand-attribute 
information . 



55 



All subjects were exposed to some basic product and 
package information about the various brands before they saw 
any ads. This was intended to stimulate real-world 
situations in which subjects typically have access to 
information sources other than advertising. Therefore the 
purpose of providing the initial information was to give 
subjects a preliminary basis for forming relative preferences 
among brands. This was considered to be especially important 
for subjects who were not to be exposed to advertising. 
Otherwise, these subjects would have no basis for choice 
(from novel brands) other than price. Afterward, subjects in 
the two advertising conditions were exposed to advertising, 
which was the second source of information about the brands. 

However, if subjects in the no-advertising condition had 
perfect recall of the initial attribute information that they 
were exposed to, they were more likely to have a well-defined 
preference ordering for the brands. In such a situation, 
advertising would be less likely to result in significant 
increases in the dispersion of brand-utilities. In short, for 
advertising to have the desired effects on the size of the 
consideration set and preferences, it was considered 
important that there should be differences among the three 
advertising conditions in terms of the information that is 
accessible . 

Effects — of — adverti sing on interbrand differences in 
preference . Advertising conveying information that produces 



56 



(real or image-based) differentiation among brands was 

postulated to have an effect on the relative overall 

preferences. Specifically, differentiating advertising is 

likely to make perceived utilities of the brands more 

variable, resulting in some being perceived to be clearly 

preferred to others. For example, it is likely that 

advertising will increase the utility difference between the 

most preferred brand and the next preferred brand. Reminder 

advertising was not expected to have significant effects on 

interbrand differences in preference. On the basis of the 

above discussion, the following predictions were made: 

P3a. Differentiating advertising will lead to a significant 

increase in mean preference ratings and the dispersion 

in preference ratings compared to the no-advertising 

condition . 

P3b. Reminder advertising will not lead to a significant 

Increase in mean preference ratings and the dispersion 

in preference ratings compared to the no-advertising 

condition . 

Effe cts of advertising on price perceptions . Consumers' 
perceptions of prices of brand alternatives are a function of 
the information available. Thus, in the absence of 
advertising, consumers are less likely to perceive 
significant differences across comparable brands. In 
contrast, when differentiating advertising conveys 
information on the attributes of the various brands, it is 



57 



likely that some will be perceived as being clearly superior 
to others. This is expected to affect price expectations for 
the various brands and also the variance in expected prices. 
Therefore the following prediction was made: 
P4. Differentiating advertising will result in higher mean 

expected prices and higher dispersion in expected prices 

compared to the no advertising condition. 

Subjects' expectations of prices were likely to be 
different across various brands. The pretest was also 
designed to help calibrate the price levels and price 
variations to be used in Experiment 3 which would directly 
test for the effects of advertising on price elasticity. It 
was desired to set normal price levels for each brand that 
would make all brands roughly equal in attractiveness, so 
that the preference orderings would be maximally sensitive to 
price discounts. This was especially crucial for a product 
category such as candy bars where consumers are not expected 
to be very price sensitive a priori. 

Effects — of advertising on interbrand variances in 

perceptions. Marketers often attempt to differentiate brands 
on certain advertised dimensions in order to alter the 
consumers' preference-ordering for the various alternatives 
available. The increased salience of certain advertised 
attributes might result in consumers using criteria provided 
in the ad for the purpose of evaluation. If differentiating 
advertising 'frames out' a relevant subset of brands based on 



58 



the attributes used for differentiation, it is likely that 
brands within a particular subset will be perceived as being 
more similar than brands in general. Brands in different 
subcategories are likely to be perceived as being more 
dissimilar when subjects are exposed to differentiating 
advertising compared to subjects in the control condition. 
This leads to the following prediction: 
P5. The variance in pairwise similarities among brands will 

be higher in the differentiating advertising condition 

than in the no advertising condition. 
Subnects 

Twenty-four subjects enrolled in an introductory 
1 

marketing course in the University of Florida participated in 

the experiment. Those who participated were given two extra 

credits in their class. 

Procedure 

The study was conducted at the behavioral laboratory 
provided by the Center for Consumer Research at the 
University of Florida. The experiment was conducted on four 
IBM PC-XT computers with enhanced graphics (EGA) monitors. 
Subjects participated in groups of one to four, with groups 
randomly assigned to between-subject s conditions. All 
subjects participated in two experimental sessions held on 
two consecutive days. The stimuli and tasks for the 
experiment are detailed in Appendix I. 



59 



Session 1. All the task instructions for the first 

session were administered using the computer. Subjects were 
first given an overview and preliminary instructions for the 
experiment (detailed in Appendix IA-1) . They were told that 
the study involved Canadian candy bars which were being 
considered for introduction into the local market. They were 
then exposed to product and package information (name, weight 
and contents) on twelve brands of Canadian candy bars one at 
a time. The information on each brand was flashed on the 
screen for 15 seconds. The brand description protocols are 
detailed in Appendix IA-2 . 

After being exposed to the brand information, subjects 
performed a recognition task, the instructions for which are 
detailed in Appendix IA-3 . The purpose of this task was to 
test whether the brand names were accessible in memory. The 
task involved correctly identifying the twelve target brands 
from a larger list of twenty four brands (which included 12 
distractor brands) . The distractor brands (listed in 
Appendix IA-4) consisted of Canadian candy bars and were 
randomly intermixed with the target brands. Subjects were 
judged to have performed satisfactorily on the recognition 
task if they were able to correctly identify all the target 
brands and had less than 75% failure rate with the 
distractors . 

Subjects who failed to satisfy the criteria for the 
recognition task were exposed to the product and package 



60 



information again before they attempted to perform the 
recognition task again. This was repeated until all subjects 
had performed satisfactorily on the test, as measured by the 
criteria listed above. The computer recorded the number of 
exposures that were required for each subject to satisfy the 
criteria for the recognition test. Subjects who finished 
earlier than the others were asked to wait quietly. When all 
subjects completed the task, they were thanked for their 
participation and asked to return for the next session. 

Session 2 . During the second session, subjects in the 
two advertising conditions were exposed to ads for the 
various brands. The experimenter showed the ads to the 
subjects in the differentiating and reminder advertising 
conditions. Subjects in the no-advertising condition did not 
see any ads . 

Advertising stimuli . The stimuli for the differentiating 
advertising condition were 30-second television commercials 
for the twelve brands of candy bars. These were provided by 
the manufacturers of the Canadian candy bars. The stimuli 
used for the reminder advertising condition were slides 
created from the television commercials, and showed the 
package and the brand name. A remote control slide projector 
was used to show the reminder ads to the subjects. The 
duration of exposure to the reminder ads for each brand was 
15 seconds. It was judged that a 30 second exposure to the 



61 



reminder ads would have produced irritation among subjects 

that would have confounded the interpretation of the results. 
After this, all subjects completed two questionnaires. 

The first questionnaire required subjects to perform the 

following tasks : 

Brand name recall task. Subjects were asked to recall 

and list the target brands of candy bars. The instructions 

for this task are detailed in Appendix IB-1. 

Attribute recall task . Subjects were provided with a 

list of candy bar names for which they had been exposed to 

information (product and package, and in some cases, 
advertising) on candy bars. They were asked to recall and 
list the attributes that describe the candy bars. The 
instructions for this task are detailed in Appendix IB-2 . 

Preference rating task . The instructions for this task 
are detailed in Appendix lB-3 . Subjects were required to 
indicate their preferences for the target brands of candy 
bars on a seven point (Extremely high— Extremely low) scale. 

Price perception i-ac,k The instructions for this task 
are detailed in Appendices IB-4 and IB-5. Subjects were 
given the list of target brand names and were asked to 
estimate the most likely, lowest likely and highest likely 
price for each brand. They were told that ' Skor ' , a brand 
with which they were familiar was currently available on the 
market for $0.55. 



62 

Similarity rating task . Subjects were asked to give 

similarity ratings for the twelve brands on a seven-point 
(highly similar — highly dissimilar) scale. The instructions 
for this task are detailed in Appendix IB-6. 
Results and Discussion 

The results of Experiment 1 are presented below. 

Eff ects of advertising on the number of brands recalled . 
It was predicted that since advertising increases the ability 
to recall brand names, subjects who were exposed to ads would 
perform better on the brand name recall task than those who 
did not see any ads. An analysis of variance on the total 
number of brand names correctly recalled across the three ad 
conditions approached significance, F(2,21) = 2.49, 2. < -10, 
(CO = .11). Moreover, there were significant differences 
between the control condition (M = 4.87) and the two 
advertising conditions (M = 7.31), F(2,22) = 4.90, p. < .05. 
There were no significant differences between the reminder (M 
= 7.12) and differentiating (M = 7.50) advertising 
conditions, F(l,21) <1 . Thus the results were in the 
predicted direction, indicating that there was a positive 
effect of advertising on the ability to recall brand names. 

Effects — of advertising on the number of attributes 

recalled . It was predicted that if differentiating 
advertising did indeed convey attribute information about the 
brands, then subjects in this condition should perform better 
on an attribute recall task compared to subjects who were not 



63 



exposed to such ads. The attributes considered were those 
that were initially presented to all subjects. An analysis 
of variance conducted on the total number of brand-attributes 
correctly recalled in the differentiating advertising 
condition (M = 6.50) and the no advertising condition (M = 
4.75) approached significance, F(l,21) =2.95, ^ < .10, (co^ = 
.04). As expected, subjects in the reminder advertising 
condition (M = 5.00) did not differ significantly from those 
in the control condition (M = 4.75) on the number of 
attributes correctly recalled, F(l,21) < 1. Thus the results 
were in the predicted direction, indicating that there was a 
simple effect of differentiating advertising on attribute 
recall scores . 

Effects — of — advertising on interbrand differences in 
preference . It was predicted that advertising that results 
in real or image-based differentiation will change the 
utilities and the variance in utilities for the various brand 
alternatives. Thus we would expect the mean preference 
ratings and the variance in ratings to be higher when 
subjects are exposed to differentiating advertising compared 
to the no advertising condition. An analysis of variance 
conducted on the mean preference ratings across the 
advertising conditions was significant, F(2,21) = 3.38, p. < 
.05, (CO = .17). There were no significant differences in 
mean ratings between the differentiating advertising 
condition (M =4.8) and the no advertising condition (M = 



64 



4.5 6), F(l,21) < 1. Although reminder ads were not expected 
to have a significant effect on preferences for the various 
brands, mean ratings were lower in this condition (M = 4.17) 
than in the control condition. 

It was predicted that differentiating advertising will 
lead to a greater variance in preference ratings in 
comparison to the no advertising condition. An analysis of 
variance performed on the standard deviation of preference 
ratings approached significance, F(2,21) = 2.16, £ < .14, (co^ 
= .09). As expected, the dispersion in brand-utilities was 
marginally greater in the differentiating advertising 
condition (M = 1.72), as compared to the control condition (M 
= 1.37), F(l,21) = 3.24, p. < .08. There were no significant 
differences between the control condition and the reminder 
advertising condition in the variance in brand-utilities, 
F.(l,21) < 1. Thus the results were in the predicted 
direction and showed that differentiating advertising did 
result in a greater spread in brand-utilities. 

Results show that the differences across advertising 
conditions were not large. it is possible that subjects 
anchored high and low ends of the preference scale with 
whatever stimuli seemed most and least preferred. Such 
response language differences across conditions could have 
thus defeated this prediction (Lynch, Chakravarti & Mitra 
1990; Upshaw 1962; Wyer 1974) . 



65 



This result provided insight on the preference measure 
that was to be used in Experiment 3. Measures of revealed 
preference, which are not likely to be susceptible to 
response language effects were used rather than rating scale 
measures of preference. 

Effects — of — advertising on price perceptjons , it was 
expected that differentiating advertising would lead to 
higher expected prices on the average compared to the control 
condition. Mean price estimates for the various brands were 
examined as a function of advertising condition. 

An analysis of variance on mean price estimates showed 
that there were no significant differences between the 
differentiating advertising condition (M = 57.34) and the no 
advertising condition (M = 53.58), F(l,21) < 1, (co^ < O). 
The dispersion of estimated prices were also predicted to be 
higher in the differentiating advertising condition in 
comparison to the no advertising condition. An analysis of 
the standard deviation of price estimates showed that the 
price dispersion was higher in the differentiating 
advertising condition (M = 8.66) than in the control 
condition (M = 5.50). But although the means were in the 
predicted direction, the differences were not significant, 
F(l,21) = 1.66 ^ < .25, (CO^ = .03) . 

Although the mean price estimates and the variance in 
price estimates showed trend in the predicted direction, the 
differences were not significant. Since there were only 



66 



eight subjects for each mean being compared, this could have 
been due to the lack of power to detect hypothesized effects. 

Estimates — of high and low prices . Besides providing 

expected most likely price estimates, subjects also reported 
highest and lowest likely prices for each brand. These 
estimates were used to calibrate high and low prices that 
were to be used in Experiment 3. The procedure that was used 
to set the discounted price for each brand is detailed in the 
next chapter. 

Effects — of — advertising on interbrand differences in 
perceptions . it was predicted that differentiating 
advertising which highlighted attributes of brands would 
result in some brands being perceived as similar to some and 
more dissimilar to others. Thus the variance in pairwise 
similarities was expected to be higher under differentiating 
advertising condition compared to the no advertising 
condition. An analysis of the variance of pairwise 
similarities showed that there were no significant 
differences across advertising conditions, F(2,21) < 1, (co^ < 
0) . Thus there is not enough evidence to conclude that 
differentiating advertising did result in brands being 
perceived as more dissimilar. 

A possible explanation for this could be that in the 
absence of advertising information, subjects used the brand 
name and attribute information (presented in the first 
session) to make judgments about the similarity of brands. 



67 



Since these were real candy bar names, many of these provided 
some insight about the different attributes possessed by the 
brands (e.g.. Crispy Crunch, Caramilk) . Therefore it is 
likely that even in the no advertising condition, subjects 
perceived brands to be quite dissimilar apriori, simply on 
the basis of their names. In comparison, brands were not 
perceived to be more different when they were advertised, 
thus resulting in no significant differences across 
advertising conditions. Also, response language effects, 
(whereby subjects anchored high and low ends of the 
similarity scale with stimuli that seemed to be most and 
least similar) could have been the reason for the inability 
to detect significant effects empirically. 

Conclusions 

The results of the first pretest were encouraging. On 
the basis of the findings it was clear that the ads that were 
being used were leading to results in the predicted 
direction. It is likely that the small sample size in the 
pretest resulted in the lack of sufficient power to detect 
hypothesized effects. In spite of this, it was clear that 
the advertising manipulation was quite effective in terms of 
bringing about expected changes in the number of brands 
recalled and the distribution of preferences. This was 
further explored in the next pretest. 



68 



Experiment 2 
In the first pretest, the analysis of variance on the 
number of brands recalled across advertising conditions was 
marginally significant (^ < .10, (O^ = .11). As predicted, 
there was a significantly positive effect on recall scores 
when subjects were exposed to advertising (M = 7.31) and 
compared to the no advertising condition (M = 4.87). 
However, it was desired to have a stronger memory 
manipulation which would result in sufficient power to detect 
significant effects on price sensitivity. Moreover, it was 
desired to reduce the average size of the consideration set 
in the no advertising condition because a proportionate 
change in consideration set size should have a larger effect 
on price elasticity if set sizes are small rather than large. 
In other words, it was important to ensure that the 
differences in recall between the advertising and control 
conditions were even greater. Therefore this pretest was 
designed to incorporate modifications in the experimental 
procedure in order for advertising to have a stronger effect 
on the size of the consideration set. In order to achieve 
this objective, three specific changes were made in the 
procedure to be used for Experiment 3. The rationale for 
these changes are detailed below. 

In the first pretest, subjects had been exposed to the 
brand information repeatedly until they met the criteria for 
satisfactory performance on the recognition task. it was 



69 



felt that an overexposure to this information was probably 
responsible for reducing the marginal effect of advertising 
in making brand names accessible in memory. Therefore it was 
necessary to limit the maximum number of times that subjects 
were exposed to this information. Accordingly, the procedure 
for this experiment was modified such that subjects could see 
the information no more than two times. 

In the first pretest, subjects in the advertising 
conditions were exposed to ads for the candy bars only during 
the second session. In order for advertising to have 
significant effects on recall it was decided that subjects in 
the advertising conditions would be exposed to the ads two 
times--once during the first session and once during the 
second session. 

It was evident from the first pretest that subjects had 
fairly high brand name recall scores in general, which made 
it difficult to detect significant improvements in recall due 
to advertising. Besides the modification in the recognition 
task to control for overexposure to brand information, a 
change was made in the procedure for this experiment. It was 
felt necessary to interfere with subjects' memory for target 
candy bar names, so that advertising might result in a 
stronger effect on recall scores. Accordingly, at the end of 
the first session, subjects were given a list of candy bar 
names to memorize. It was hoped that names from the same 
product category would interfere with their memory for the 



70 



target brand names so that advertising might have a 

J significant effect on recall scores. 

Ill 

Subjects 

Twenty-four subjects enrolled in an introductory 
marketing course in the University of Florida participated in 
the experiment. Participants were given two extra credits in 
their class . 
Procedure 

All subjects participated in two experimental sessions. 

I 
In the first session they were exposed to product and package ■" 

information (via the computer) on the brands and were given 

the modified recognition task to ensure that the brand names 

were available in memory. This implied that if even after 

two exposures to the brand information subjects did not 

perform satisfactorily on the recognition task, they went on [ 

to the next stage of the experiment. ' 

After this, subjects in the differentiating advertising 

condition saw twelve ads, while subjects in the reminder 

advertising condition saw twelve reminder ads for the candy 

bars. The modified instructions to the subjects at this « 

stage of the experiment are detailed in Appendix IC-1 . 

Subjects in the no advertising condition did not see any ads 

and therefore went on to the next task. The advertising 

stimuli that were used for this experiment were the same as 

those used in the first pretest . 



71 



After this, subjects were given three minutes to 
memorize a list of twenty-four candy bar names which included 
twelve American and twelve European brands available in the 
Gainesville market. The instructions and stimuli for this 
task are listed in Appendix IC-2 and IC-3 respectively. When 
they had completed this task they were thanked and asked to 
come back for the next session. 

In the second session, all subjects completed the brand 
name recall task where their objective was to recall and list 
as many brand names as they could. Before they performed 
this task, subjects in the advertising conditions were 
exposed to ads for the candy bars. These ads were the same 
as those that they had seen in the first session. Subjects 
in the no advertising condition did not see any ads and 
completed the questionnaire directly. 
Results and Discussion 

It was predicted that advertising will have a 
significant effect on the total number of brands recalled. 
An analysis of variance across the three advertising 
conditions was significant, F(2,21) = 18.59, u < .0001, (co^ 
= .59), Subjects in the two advertising conditions recalled 
a significantly greater number of brands than subjects in the 
control condition (M = 8.43 and M = 3.00 respectively), 
F(2,21) = 38.57, ^ < .0001. As in Experiment 1, there were 
no significant differences in recall scores between the 



72 



reminder (M = 8.25) and differentiating advertising (M = 
8.62) conditions, F(l,21) <1. 

On the basis of these findings it was possible to 
demonstrate with this independent sample of subjects that 
advertising could indeed have significant effects on the 
memory for brand names. Therefore in choice environments in 
which memory for brand names is important such effects can be 
expected to prevail. if the ability to recall brand names is 
the only determinant of the number of brands considered, we 
can expect consideration set sizes across different 
advertising conditions to follow the same pattern as brand 
name recall scores. Thus this pretest provided a benchmark 
against which to compare and interpret the results on 
consideration set size. The next chapter details the 
experiment that tests for advertising effects on price 
sensitivity, mediated by the preferences and the size of the 
consideration set. 



CHAPTER V 

AN EMPIRICAL INVESTIGATION OF THE IMPACT OF ADVERTISING 
ON PRICE ELASTICITY: RESEARCH METHODOLOGY 

Chapter Overview 
The propositions developed in Chapter III concerned how 
advertising could influence consumer price sensitivity. The 
relationship between these two variables was conceptualized 
as being mediated by two key constructs — the size of the 
consideration set and the dispersion of brand-utilities. It 
was argued that advertising could have either a positive or a 
negative effect on the number of brands considered at the 
time of choice. Hence, by increasing or decreasing the size 
of the consideration set, advertising could have a positive 
or negative effect on price sensitivity. Besides affecting 
the size of the consideration set, advertising could also 
increase the spread of utilities of brands and result in some 
brands being perceived as being clearly superior to others. 
This would lead to a decrease in price sensitivity. 

In this study on the economic effects of advertising, 
the content of advertising (whether reminder or 
differentiating) and the decision environment (whether 
stimulus-based or memory-based) were hypothesized to moderate 
the relationship between advertising and price sensitivity. 
It was predicted that in memory-based decision environments, 



73 



74 



advertising is likely to influence the ability to recall 
brands and also the spread of brand-utilities. This was 
postulated to have implications for the number of brands 
considered for choice, the direction being determined by the 
relative strength of these two factors. In stimulus-based 
choice environments, where the ability to recall a brand is 
not an important factor for it to be a candidate in the 
consideration set, the effect of advertising is primarily 
mediated by the dispersion of brand-utilities. 

This chapter details the methodology employed in 
Experiment 3 to test these hypotheses . The remainder of this 
chapter is divided into three main sections which describe 
(1) the subjects and stimuli used for the study, (2) the 
rationale for the factors employed, (3) the dependent 
variables, (4) the stimulus design, and (5) the experimental 
procedure that was employed. 

Subjects 
One hundred and ninety-eight students enrolled in the 
Introductory Marketing course at the University of Florida 
participated in the experiment. Subjects received two extra 
credits in their class and $3.00 in candy of their choice for 
participating in the study. Subjects had to participate in 
both sessions in order to successfully complete the 
experiment. Seven subjects failed to complete the study and 
their data were not used for the analysis. 



75 



Stimuli 

The product category used for this study was candy bars. 
Since the experiment used real products and real ads for 
these products it was necessary that the existing advertising 
for the products had certain desirable properties. First, 
for the purpose of this experiment it was necessary to have a 
product category with which subjects were relatively familiar 
so that they could make trade-offs between prices and other 
attributes relatively easily. Second, it was necessary that 
it be a category in which there exists attribute-based 
differentiation among brands so that it is likely that the 
advertising could highlight such differences. In both these 
respects the product class of candy bars was considered 
appropriate for the study. 

The study attempted to examine the effect of advertising 
on price sensitivity mediated by preferences and the size of 
the consumers' consideration set. In order to be able to 
draw unambiguous conclusions about the effect of advertising 
on these mediating variables it was also crucial that the 
subject population not have prior familiarity with the brands 
and the ads that were to be used in the experiment. 
Therefore the stimuli for the study were twelve Canadian 
candy bars . 



76 



Experimental Factors 
This section details the experimental factors that were 
manipulated in the study. As noted earlier, the study 
employed a design in which advertising condition and decision 
environment were between subjects factors and price of twelve 
brands was manipulated within subjects . 
Advertising Condition 

Advertising Condition was manipulated at three levels in 
order to gauge the effects of advertising content on price 
elasticity. Subjects in the differentiating advertising 
condition saw ads for all twelve candy bars. These ads 
provided information on brand attributes and were predicted 
to change subjects' relative preference for the brands 
compared to a situation where none of the brands were 
advertised. Subjects in the reminder ad condition saw ads 
which provided information on the package and brand name, 
similar to some print or billboard advertising. These ads 
were predicted to increase the likelihood of inclusion of the 
advertised brands in the consideration set. Subjects in the 
no advertising condition served as a control group and were 
not exposed to any ads for the candy bars. 

An alternative manipulation would have been to advertise 
a subset of brands instead of all of them. This would have 
been closer to real-world situations where advertised and 
unadvertised brands exist simultaneously. Also, this 
manipulation could probably have provided better insights on 



77 



the signalling properties of advertising and how advertising 
affects preferences. But this might have led to demand 
artifacts, especially in a rating/choice situation. Also 
advertising for a few brands was likely to lead to inhibition 
effects (Alba & Chattyopadhyay 1986) . It was felt that the 
power of detecting any of the hypothesized effects was also 
likely to be greater if all the brands were advertised. 
Decision Environment 

Subjects went on shopping trips in which they made 
choices among the candy bars. In order to gauge the effects 
of consideration set size on price elasticity and the effects 
of advertising on consideration set size, the decision 
environment was manipulated at two levels. Subjects in the 
stimulus-based choice condition were provided with a list of 
all twelve brands and could therefore use this list to decide 
which brands to examine for price information and purchase. 
This was intended to reflect real-world situations in which 
the aisle display or a salesperson reminds the consumer of 
the set of brands that might actually be purchased. 

Subjects in the memory-based choice condition did not 
have such a list available to them. These subjects could 
only search for price information on brands that they were 
able to recall on their own. Therefore they could only 
purchase brands that they could retrieve from memory. This 
manipulation was intended to reflect real-world decision 
situations in which the consideration set is typically 



78 



generated from memory. Once they could recall a brand name 
correctly, subjects had access to the same information that 
was available to subjects in the stimulus-based decision 
environment, because attribute information was memory-based 
for both groups . 

Thus the only difference between these two groups was 
whether they had to recall the brand names or had to access 
them from memory. In the absence of any advertising for the 
brands, the differences between groups in the two decision 
environments could be attributed purely to memory factors. 
This provides better insights on the cuing properties of 
advertising and how advertising influences the size of the 
consideration set. 

Dependent Variables 
For each subject, the following dependent measures were 
collected: 

1. measures of average consideration set size; 

2. measures of dispersion — the proportion of budget 
allocated to the top brand, the number of brands ever 
considered, the standard deviation of the proportion of 
budget allocated to the various brands, the standard 
deviation of the frequency of choosing the various 
brands ; 

3. price elasticity of each brand. 



79 



Stimulus Design for Price Manipulation 

The prices of the twelve brands were manipulated within- 

subjects at two levels (high or low) independent of each 

other. It was determined that sixteen shopping trips were 

required in order for prices to vary orthogonally for the 

twelve brands. Whether a particular brand was at its high or 

low price was varied within subjects in a l/256th fraction of 

a 2l2th within-subjects design. Sixteen brand-price profiles 

were constructed such that across these profiles each brand 

was at its high price in eight brand price profiles, and at 

its low price for the other eight with the prices of the 

brands uncorrelated with each other across trials. The 

brand-price profiles used for the experiment are detailed in 

Appendix IIA-1 . The order in which subjects saw these brand 

price profiles for the sixteen shopping trips was 

counterbalanced across subjects in a 16X15 diagram-balanced 

Latin Square design. Thus each subject saw the brand-price 

profiles in a particular sequence. There were sixteen such 

between-subjects sequences to which subjects were assigned. 

The sixteen level sequence factor described above 
simultaneously counterbalanced order of presentation of the 
sixteen price profiles and the order of presentation of 
brand-triples in advertising. The latter will be explained 
later . 

Subjects were told that the prices of the twelve brands 
were to vary from one shopping trip to another. At a 



80 



particular point in time, a given brand was at its high price 
for half the subjects and at its low price for the other 
half. The price manipulation was used to compute independent 
measures of self price elasticities at the brand level. 

The high prices were calibrated from the average highest 
likely price estimates (which are detailed in Appendix IIA-2) 
that were obtained from the control group of subjects in 
Experiment 1. in order to detect effects on price 
sensitivity it was necessary that the price differential was 
sufficient to cause a change in purchase patterns. The mean 
lowest likely price estimates obtained from the control group 
in Experiment 1 were used to calibrate the low prices for the 
brands. The figures were rounded off to the nearest 5 cents 
to arrive at the high prices for the various brands. The low 
price used in the experiment was 25 cents lower than the high 
price, which worked out to an average 34% discount. The 
actual high and low prices that were used for the twelve 
brands are detailed in Appendix IIA-3. 

Procedure 
To test the hypotheses, a computer-based shopping 
simulation was developed and employed. The study was 
conducted at the behavioral laboratory provided by the 
Center for Consumer Research at the University of Florida. 
The experiment was conducted on four IBM PC-XT computers with 
enhanced graphics (EGA) monitors. All subjects participated 



in two experimental sessions held on two consecutive days. 
All experimental instructions were administered using the 
computer, thus limiting the experimenter's role during the 
experiment to exposing the subjects to advertising, answering 
questions and handling minor administrative details. This 
reduced experimenter-induced biases and ensured that all 
subjects received the same task instructions. All data were 
collected by the computer eliminating the chances of 
transcription or keypunching errors. 

Subjects participated in batches of one to four each. 
Upon entering the room, each subject was randomly assigned to 
one of the four sequences nested in an Order. Eight subjects 
completed the experiment for each of the 24 between-sub jects 
cells (described in Appendix IIB-2) . 
Session 1 

Subjects were first given an overview and preliminary 
instructions for the experiment. They were told that the 
study involved Canadian candy bars which were being 
considered for introduction into the local market. They were 
then exposed to product and package information on twelve 
brands of Canadian candy bars one at a time. The information 
on each brand was flashed on the screen for 15 seconds. 

The sequence in which this information was presented to 
subjects was counterbalanced. The twelve brands were divided 
into four triples of three brands each. These are detailed 
in Appendix IIB-1. Thus the order of presentation of 



82 



information on these twelve brands can be conceptualized as 
the order of presentation of information on four triples. 
The sixteen sequences corresponding to presentation order for 
price profiles were divided into four groups (Orders) with 
four sequences nested in each (detailed in llB-2). The 
position of these brand-triples in the string of twelve was 
varied across the four groups in a 4X4 Latin square design 
(detailed in Appendix llB-3) . Therefore the position of each 
particular triple was counterbalanced across subjects. 

Recognition task . After being exposed to the brand 
information, subjects performed a recognition task, the 
purpose of which was to test whether the brand names were in 
memory. The task involved correctly identifying the twelve 
target brands from a larger list of twenty-four Canadian 
candy bar names randomly intermixed with the target brands. 
Subjects who failed to satisfy the criteria for the 
recognition task were exposed to the product and package 
information once more before they attempted to perform the 
recognition task again. If after the second exposure to the 
brand information they did not perform satisfactorily on the 
recognition test, they went on to the next stage of the 
experiment . The computer recorded the number of exposures 
required for each subject to satisfy the criteria for the 
recognition test. Subjects who finished earlier than the 
others were asked to wait quietly. 



83 



Advertising exposure manipulation . Subjects were now 
exposed to ads for the candy bars. Subjects in the 
differentiating advertising condition saw twelve 
differentiating ads, while subjects in the reminder 
advertising condition saw twelve reminder ads for the candy 
bars. Subjects in the no advertising condition did not see 
any ads and therefore went on to the next task. 

The sequence in which these ads were presented to 
subjects was counterbalanced. The ads for the four brand- 
triples (described in Appendix IIB-1) were divided into four 
triples of three ads each. The position of ads corresponding 
to these brand-triples was varied across the four Orders in a 
4X4 Latin square design. This is detailed in Appendix IIB-4 . 

The stimuli used in this study were those that were used 
in the pretests. The stimuli for the differentiating 
advertising condition were 30-second television commercials 
for the twelve brands of candy bars. The stimuli used for 
the reminder advertising condition were slides taken from the 
television commercials, and showed the package and the brand 
name. A remote control slide projector was used to show the 
reminder ads to the subjects. The duration of exposure to 
the reminder ads for each brand was 15 seconds. 

Interference task . After this, subjects were given three 
minutes to memorize a list of twenty-four candy bar names 
which included twelve American and twelve European brands 
available in the Gainesville market. This was the same list 



84 



that was used for Experiment 2. When they had completed this 
task they were thanked and told to come back for the next 
session . 
Session 2 

The introduction and preliminary instructions for the 
second session are detailed in Appendix IIC-1 and IIC-2. 

Shopping trips . Subjects were first told that they would 
be going on 16 different shopping trips for the Canadian 
candy bars which they had seen the day before. They were 
told that they had $3.00 to spend on each shopping trip. 
They were also told that they would actually receive the 
candy bars that they purchased on one of these shopping 
trips. They did not know in advance which of these shopping 
trips would be involved in their actual purchase — this was to 
be randomly determined. 

Subjects were informed that prices of the candy bars 
would be varying from one shopping trip to another, so that a 
brand that was at a low price on one of the shopping trips 
would not necessarily be at that price on the next trip. 
They were told that on any particular shopping trip they 
could ask for price information on as many or as few brands 
as they wanted. 

Practi ce trips . Before shopping for candy bars, subjects 
went on four practice shopping trips. The purpose of these 
practice trips was to familiarize subjects with shopping on 
the computer and to ensure that they knew how to search for 



85 



price information on the brands that they liked. It was 
hoped that these practice trips would demonstrate to them 
that prices were indeed varying from trip to trip. The 
introduction to the practice trips are detailed in Appendix 
IIC-3. 

The product category of soaps was used for the practice 
trips. It was desirable that the product category used for 
the practice trips was one in which there are not many 
perceived differences between brands so that price would be 
an important factor in decision making. Since candy bars 
were to be used for the main study, and consumers are 
believed not to be very price sensitive in that category, it 
was important to devise ways that would increase the weight 
given to price, and thus increase the power of the experiment 
to detect effects. In choosing a product category it was also 
necessary to ensure that it was not one that would make the 
price differences for candy bars seem insignificant. 
Therefore six familiar brands of soaps were used for the 
practice trips. 

The prices of soaps were varied independently of each 
other. Four brand-price profiles were constructed with each 
brand being at its high price for two profiles and at its low 
price for two profiles. 

Subjects went on four practice trips. On each trip, the 
computer made available the list of six brand names for which 
they could search for price information. When subjects 



86 



requested information by typing in the brand name, the 
computer responded with the price for that particular brand. 
Subjects could search for information on as few or as many 
brands as they liked. When they specified that they did not 
require further information they were instructed to allocate 
their budget of $3.00 among soaps of their choice. After 
each purchase, the computer tallied the remainder of the 
budget that was still available. After completing one 
particular practice shopping trip, subjects went on to the 
next trip. The instructions for the practice trips are 
detailed in Appendix llC-4. 

Ad vertising Exposure #1 . After the practice shopping 
trips, subjects in the advertising conditions were exposed to 
ads for the candy bars. Subjects in the no advertising 
condition did not see any ads and therefore went on to the 
next task. 

A similar procedure was used for the advertising 
exposure manipulation for the second session as in the first 
session. The instructions before the exposure to advertising 
is detailed in Appendix IID-1. The same 4X4 Latin Square 
design (detailed in IlB-4) was used for this purpose, thus 
resulting in subjects seeing the ads in the same sequence in 
this session compared to the first session. 

Shopping trips. After this, subjects went on 16 shopping 
trips for candy bars. Subjects in the stimulus-based choice 
condition were given a list of the twelve candy bars and they 



typed brand names for which they wanted price information. 
The instructions for this condition are detailed in Appendix 
IlD-2 . Subjects in the memory-based choice condition did not 
have such a list available to them, and had to retrieve brand 
names from memory. The introduction and instructions for the 
memory-based condition are detailed in Appendix IID-3. 

Subjects could search for price information on as few or 
as many brands as they wished. When subjects requested 
information by typing in the brand name, the computer 
responded with the price for that particular brand. The 
prices at which a subject saw a particular brand on a 
particular trip depended on the sequence number that s/he was 
assigned at the beginning of the experimental session. As 
noted earlier, across the sixteen shopping trips a subject 
saw brand-price profiles in one of sixteen possible 
sequences . 

After they had finished searching for information on 
prices, they were instructed to allocate their budget of 
$3.00 among candy bars of their choice. After each purchase, 
the computer computed the remainder of the budget that was 
still available. After completing one particular trip, 
subjects went on to the next trip. 

Advertising Exposure #2 . Since the shopping trips were 
very similar to each other, it was felt that the positive 
effects of advertising on consideration set size would not be 
as strong with only one exposure to advertising. Thus there 



would be a tendency for only a few brands to be included in 
the consideration set. In order to increase the power to 
detect significant effects, subjects in the two advertising 
conditions were exposed to ads for the candy bars once more 
after completing eight shopping trips. The sequence in which 
they saw the ads for the brands during the second exposure 
was the same as in the first exposure. 

After this, subjects went on the remaining shopping 
trips for candy bars, with prices varying from one trip to 
another. In all these trips, subjects allocated $3.00 among 
candy of their choice. Subjects who had finished earlier 
than the others were instructed to wait quietly for the other 
participants to complete the experiment. 

After they had completed the sixteen shopping trips, 
subjects were asked what strategies they had used in making 
their choices. They were also told that the computer would 
randomly select one of the 16 shopping trips for each of 
them and that they would get the candy bars that they had 
purchased on that shopping occasion. They were thanked for 
their participation and informed that they could collect 
their candy bars at the end of the semester. 

The results of this study are presented in the next 
chapter . 



CHAPTER VI 

AN EMPIRICAL INVESTIGATION OF THE IMPACT OF ADVERTISING 
ON PRICE ELASTICITY: RESULTS 



Chapter Overview 

This chapter presents the results of Experiment 3 in two 
parts. The results pertaining to the average size of the 
consideration set and the dispersion of brand-utilities are 
first presented. Subsequently the findings corresponding to 
price sensitivity are discussed. 

The purpose of this study was to provide a better 
understanding of the effects of advertising on price 
elasticity by isolating the causal mechanisms underlying 
these effects. It was proposed that the relationship between 
advertising and price sensitivity is mediated by two basic 
theoretical construct s--consideration set size and the 
dispersion of brand-utilities. Therefore, before specific 
hypotheses linking advertising with price sensitivity were 
tested, it was necessary to verify that advertising did have 
the predicted effects on these two mediating variables. 

Results 
Consideration Set Size 

Consideration set was measured at the individual subject 
level by the number of brands about which the subject asked 
for price information on a given shopping trip averaged 

89 



90 



across the 16 shopping trips. The data were analyzed to test 
for the effects of Decision Environment, Advertising 
Condition and Sequence of Presentation on the number of 
brands considered per shopping trip. The results from an 
analysis of variance are presented in Tables VI-1 and Vl-2 . 



TABLE VI -1 



ANALYSIS OF VARIANCE 



SOURCE 


TYPE III SS 


DF 


F VALUE 


P > F 


ENVIRONMENT 


61 


45314514 


1 


15.82 





0001 


ADVERTISING 


52 


50655754 


2 


6.76 





0018 


ENVIRON*AD 


98 


99685669 


2 


12.74 





0001 


SEQ 


42 


59451248 


15 


0.73 





7476 


ENV*SEQ 


57 


52642877 


15 


0.99 





4748 


AD*SEQ 


50 


40397701 


30 


0.43 





9947 


ENV*AD*SEQ 


113 


16476440 


30 


0.97 





5197 


SS/ENV,AD, SEQ 


369 


49648233 


95 









TABLE VI -2 
MEAN CONSIDERATION SET SIZE 

ENVIRONMENT 



ADVERTISING 
NONE 

REMINDER AD 
DIFFERENTIATING 



STIMULUS 
BASED 



5.24 
4.72 
4.56 



MEMORY 
BASED 



2.26 

5.25 
3.59 



91 



The results indicate that there were significant main 
effects of Advertising Condition and Decision Environment and 
a significant Advertising Condition X Decision Environment 
interaction. To examine more closely the nature of 
advertising effects under the two decision environments, 
analyses of planned comparisons were carried out. 

Recall from Chapter III that it had been predicted that 
in the absence of advertising, consideration sets will be 
smaller in the memory-based choice condition than in the 
stimulus-based choice condition. This is because in the 
stimulus-based condition, brand names are readily available, 
and therefore it is not necessary to retrieve them from 
memory. This prediction was supported by a significant main 
effect of decision environment. The size of the 
consideration set was significantly smaller in the memory- 
based condition (M = 2.26) compared to stimulus-based 
condition (M = 5.24) , F(l,95) =36.44, ^ < .001, (CO^ = .06) . 

Consideration sets in s t imu lu s -bas ed decision 

environments . It was predicted that under stimulus-based 
choice environments, the primary effect of advertising on the 
size of the consideration set will be through its effect on 
the dispersion of preferences. Specifically, it was proposed 
that reminder advertising will not have a significant effect 
on the size of the consideration set, while differentiating 
advertising will increase the interbrand variance of 
utilities and thereby reduce the size of the set. The 



92 



findings are presented in Table VI-3 and VI-4 . Results 
showed that there were no significant differences in the size 
of the consideration set between the reminder advertising 
condition (M = 4.72) and the no advertising condition (M = 
5.25), F(l,95) < 1. This finding lends support for the 
prediction that reminder advertising will not have an effect 
on the number of brands considered when the environment 
provides the brand names. This is because the recall cues 
provided by such ads are not crucial for the retrieval of 
brand names for consideration. However, even though the mean 
number of brands considered was less in the differentiating 
advertising condition (M = 4.56) than in the no advertising 
condition (M = 5.24), this contrast was not significant, 
Z{1,95) = 1.10. Therefore there is no support for the 
prediction that differentiating advertising will lead to a 
reduction in the size of the consideration set. 

TABLE VT-.3 

CONSIDERATION SET IN STIMULUS-BASED ENVIRONMENT 
REMINDER VERSUS NO ADVERTISING CONDITIONS 

SOURCE TYPE III SS DF 

ADVERTISING 4.28619385 
SEQ 56.14105225 

AD*SEQ 39.95892334 

SS/ENV,AD,SEQ 369.49648233 



DF 


F VALUE 


P>F 


1 


1.10 


0.251 


15 


0.96 


0.687 


15 


0.68 


0.885 


95 







93 



DF 


F VALUE 


P>F 


1 


1.91 


0.103 


15 


0.86 


0.750 


15 


0.94 


0.909 


95 







TABLE VI-4 

CONSIDERATION SET IN STIMULUS -BASED ENVIRONMENT 
DIFFERENTIATING VERSUS NO ADVERTISING CONDITIONS 



SOURCE TYPE III SS 

ADVERTISING 7.43414307 
SEQ 50.42962646 

AD*SEQ 55.20159912 

SS/ENV,AD,SEQ 369.49648233 



Consideration sets in memory-based decision 

environments . It was predicted that in memory-based 
environments, there will be a significant simple effect of 
advertising on the size of the consideration set. The 
evidence provides strong support for this prediction of 
significant differences across advertising conditions, 
F(2,95) = 18.47, ^ < .001, (W^ = .05). In order to examine 

more closely the nature of effects, analyses of planned 
contrasts were carried out. Results (presented in Table VI- 
5) showed that there was a significant positive effect of 
reminder advertising on consideration set size, F(l,95) = 
36.78, E. < .001. The mean size of the consideration set was 
5.25 in the reminder advertising condition compared to 2.26 
in the no advertising condition. This lent support to the 
prediction that in an environment where brand names have to 
be retrieved from memory, the average size of the 



94 



consideration set will be larger under reminder advertising 
conditions compared to the control condition. 

It was expected that under memory-based conditions, the 
size of the consideration set will be larger in the no 
advertising condition than in the differentiating advertising 
condition. Contrary to this prediction, the number of brands 
considered were significantly smaller in the control 
condition (M = 2.26) than in the differentiating advertising 
condition (M = 3.67), F(l,95) = 6.9, ^ < .01. This result 
(presented in Table VI-6) lends further support to the 
assumption that the size of the consideration set is a 
positive function of the ability and motivation to retrieve a 
number of brands from memory. In the absence of advertising, 
the consumer has to rely on memory to recall brands for 
consideration. In comparison, advertising increases the 
ability to recall brands and therefore increases the size of 
the consideration set. 

The results of Experiment 2 provided evidence that 
advertising increases the ability to recall brand names and 
that both reminder and differentiating advertising are 
equivalent in this respect. If consideration set size is a 
function of ability alone and is unaffected by the motivation 
to recall brand names from memory, then we would expect no 
differences in the effects of the two types of advertising on 
the size of the consideration set. However, results 
(presented in Table VI-7) showed that the average size of the 



95 



TABLE VI -5 

CONSIDERATION SET IN MEMORY-BASED ENVIRONMENT 
REMINDER VERSUS NO ADVERTISING CONDITIONS 



SOURCE 



TYPE III SS 



DF 



F VALUE 



P>F 



ADVERTISING 142.87719727 
SEQ 37.16577148 

AD*SEQ 47.53295898 

SS/ENV,AD,SEQ 369.49648233 



1 


36.78 


0.001 


15 


0.63 


0.717 


15 


0.81 


0.514 


95 







TABLE VI -6 

CONSIDERATION SET IN MEMORY-BASED ENVIRONMENT 
DIFFERENTIATING VERSUS NO ADVERTISING CONDITIONS 



SOURCE 



TYPE III SS 



DF 



F VALUE 



P>F 



ADVERTISING 2 6 


20667614 


1 


6.99 





010 


SEQ 4 9 


52470911 


15 


0.84 





750 


AD*SEQ 29 


13850565 


15 


0.49 





909 


SS/ENV,AD,SEQ 3 69 


49648233 


95 









TABLE VI -7 

CONSIDERATION SET IN MEMORY-BASED ENVIRONMENT 
DIFFERENTIATING VERSUS REMINDER ADVERTISING CONDITIONS 



SOURCE 



TYPE III SS DF 



F VALUE 



P>F 



ADVERTISING 

SEQ 

AD* SEQ 



42. 96306818 
53.16983876 
52.12819980 



SS/ENV,AD,SEQ 369.49648233 



1 
15 
15 
95 



11.06 
0.91 
0.92 



0.010 
0.698 
0.684 



96 



consideration set was significantly smaller under the 
differentiating advertising condition (M = 3.59) compared to 
the reminder advertising condition (M = 5.25), £.(1,95) 
= 11.06, :g_ < .01. This suggests differentiating advertising, 
which provided more information on the relative utilities of 
the brands, resulted in reduced motivation to consider less 
favored options. The significant differences in 
consideration set size across these two advertising 
conditions suggest that the motivation to recall brand names 
did have a role in determining the size of the consideration 
set. Thus, though both types of advertising increased the 
ability to recall brand names from memory, they 
differentially affected the motivation to retrieve brand 
names in a prepurchase situation under which the objective 
was to search for prices. In the latter case, only the 
preferred brands were likely to be retrieved. An analysis of 
the effects of the two types of advertising on the preference 
lends further support for this prediction. 
Measures of Tn terbrand Variance in Preference 

Four behavioral measures of preference were used for the 
purposes of examining the effect of differentiating 
advertising on the dispersion of brand-utilities. These were 
determined from subjects' purchases of various brands under 
the different price conditions that they faced. For the 
individual subject, the following measures of preference (as 
revealed by their purchases) were computed. 



97 



Standard — deviati on of choice f requenr-.i pr . For each 
subject, the number of times each brand was bought was 
computed across shopping trips. The standard deviation of 
the twelve brands' choice frequencies was taken as a measure 
of dispersion of brand-utilities. It was expected that in 
the absence of any perceived utility differences across 
brands, each brand has approximately equal probability of 
being chosen. Thus the variance in the choice frequencies 
for the various brands were expected to be small in the no 
advertising condition. Compared to this, the variance in 
choice frequency was expected to be greater in the 
differentiating advertising condition, in as much as the 
information from such advertising was likely to make some 
brands preferable to others. Reminder advertising was not 
expected to significantly affect the variance in choice 
frequencies . 

Standard deviation of share of the budget . This is 

similar to the above measure but looks at the dispersion in 
the percentage of actual dollar expenditures. For each 
subject, the share of each brand in the total budget across 
the shopping trips was computed. The standard deviation of 
the percentage shares of the 12 brands was taken as a measure 
of dispersion of brand-utilities. It was expected that if 
there are no significant differences in perceived utility 
across brands, then each brand has a constant probability of 
being chosen. Since all the brands were in the same price 



98 



range, on an average there should not be any significant 
differences in the percentage of budget allocated to the 12 
brands. Therefore, the variance in the percentage of budget 
allocated to the various brands was expected to be small in 
the absence of advertising. When differentiating advertising 
provides attribute information that leads to some brands 
being perceived to be clearly superior to others, the 
variance is expected to increase. Reminder advertising is 
not expected to increase the dispersion in the proportion of 
budget allocated to the various brands. 

Share of budget allocated to top brand . For the 

individual subject, the percentage of budget allocated to the 
top brand was computed. It was expected that under the no 
advertising condition where there were no clearly superior 
alternatives, the share of the budget allocated to the top 
brand would be lower compared to the differentiating 
advertising condition. In the latter case, it was predicted 
that the share of the top brand would be higher because there 
would be a greater tendency on the part of subjects to buy a 
favored brand irrespective of the price. 

Number of brands ever chosen . For each subject, the 

number of brands ever purchased across the 16 shopping trips 
was computed. It was expected that in the absence of 
differentiating information on attributes of the brands, 
consumers would be relatively indifferent between options. 
Therefore it could be expected that in the no advertising 



99 



condition, subjects would make purchases on the basis of 
price. This would increase the number of brands that are 
tried, because across different trials, different brands are 
likely to be lowest in price. In contrast, if 
differentiating advertising leads to some brands being 
perceived to offer greater utility to subjects, it is less 
likely that they would purchase clearly inferior brands only 
on the basis of its price advantage. Thus differentiating 
advertising would reduce the number of brands chosen. 
Reminder advertising was not expected to have any significant 
effects on preference and therefore on the total number of 
brands chosen. 

Results — of effect s of advertising on measures of 

Interbrand variance in preferences . Before a discussion of 
the results it is important to note that these behavioral 
measures of preference were based on actual purchases made by 
subjects. The experimental design was such that subjects 
could only purchase what they could recall. Therefore, to 
examine the effect of advertising on utilities for the 
various brands it is necessary to have measures of revealed 
preference that are not confounded with the ability to 
recall. Hence, the results of the effect of advertising on 
the variance of all four measures of interbrand variances in 
preference will be analyzed for the stimulus-based 
environment only. 



100 



In the memory-based decision environment where the size 
of the consideration set itself is determined by recall 
factors, conclusions about preferences are more difficult. 
For example, a subject might be unable to recall brand names 
in the no advertising condition and this would prevent 
him/her from purchasing those brands. In such a case, the 
variance in the share of the budget allocated to the 12 
brands would be high, not because of preference for the 
brands bought but rather because of the inability to recall 
other brands for consideration and purchase. 

The means for the four measures of interbrand dispersion 
in preference are presented in Table VI-8 below. 

TABLE VI -8 
MEAN DISPERSION OF BRAND-UTILITIES 

NONE REM DIFF 

DISPERSION IN CHOICE 

FREQUENCY 4.2480 4.9334 4.9981 

DISPERSION OF SHARE 

OF BUDGET 0.0809 0.0940 0.0986 

SHARE OF BUDGET TO TOP 

BRAND 0.2 605 0.2 953 0.2 94 6 

NUMBER OF BRANDS EVER 

CHOSEN 9.7812 8.9375 8.4062 



101 



Effects of differentiating advertising . An MAN OVA 

analysis was conducted on the data for the four preference 
measures to test for effects of differentiating advertising 
on measures of interbrand variance in utilities. Results 
showed that there were significant differences between the no 
advertising condition and the differentiating advertising 
condition, F(4,29) = 2.95, ^^ < .03. This supports the 
prediction that differentiating advertising did have an 
effect on variance of interbrand preferences . 

Univariate tests also show that differentiating 
advertising did have an effect on preferences. It is evident 
that differentiating advertising did have a significant 
effect on the dispersion in choice frequencies. Table VI-9 
presents the results from the analysis of variance which 
indicates that the difference in the dispersion of choice 
frequencies between the no advertising condition (M = 4.25) 
and the differentiating advertising condition (M = 4.99) 
approached significance, F(l,62) = 2.78, ^ < .10 (CO^ = .03). 
This finding supports the prediction that differentiating 
advertising will increase the standard deviation of choice 
frequencies . 

The results presented in Table VI-10 also support the 
prediction that differentiating advertising leads to a 



significant increase in the dispersion of proportions, 
F(l,62) = 4.31, £ < .04, (co^ = .04). The standard deviation 

in the percentage of budget allocated to the various brands 



102 



was significantly greater under differentiating advertising 
(M = .10) than under the no advertising condition (M = .08) . 

Although differentiating advertising did have the 
predicted effect on the proportions allocated to the various 
brands, a similar pattern of results is not found for the 
share of the budget allocated to the top brand. Analysis of 
variance results presented in Table VI-11 show no significant 
differences in the share of the budget allocated to the top 
brand, £(1,62) = 1.34, ^ < .25, (CO^ = .01). Thus the 

findings do not support the prediction that the share 
allocated to the top brand will increase under the 
differentiating advertising condition in comparison to the no 
advertising condition. A possible reason for this is that 
advertising led to a subset of brands being considered as 
acceptable, without any clear top choice. As a result, it is 
likely that if the top brand is at its high price, it was 
bought in lesser quantities as subjects switched to one of 
the other well-liked brands. 

Table VI-12 presents the results for the analysis of 
variance on the number of brands ever chosen under the 
differentiating advertising (M = 8.40) and the no advertising 
(M = 9.78) conditions. It is evident from the findings that 
differentiating advertising led to a significant decrease in 
the number of brands ever purchased, F(l,62) = 5.68, ^ < -01, 
(CO = .08) . Thus the results support the prediction that 
differentiating advertising will reduce the number of brands 



103 



TABLE VI -9 

DISPERSION IN CHOICE FREQUENCIES 
DIFFERENTIATING VERSUS NO ADVERTISING CONDITIONS 



SOURCE 


TYPE III SS 


DF 


F VALUE 


P > F 


ADVERTISING 


8.83382541 


1 


2.78 


0.1053 


SEQ 


26.26884830 


15 


0.55 


0.8899 


AD'^SEQ 


40.24464510 


15 


0.84 


0.62 63 


SS/Ad,SEQ 


101.76058461 


32 







TABLE VT-in 

DISPERSION IN PROPORTION OF BUDGET 
DIFFERENTIATING VERSUS NO ADVERTISING CONDITIONS 



SOURCE 



TYPE III SS 



DF 



F VALUE 



P > F 



ADVERTISING 
SEQ 

AD*SEQ 
SS/AD,SEQ 



0.00499094 1 

0.01234201 15 

0.02090818 15 

0.03703894 32 



4.31 
0.71 
1.20 



0.0460 
0.7555 
0.3180 



TABLE VI- 11 

SHARE OF BUDGET ALLOCATED TO TOP BRAND 
DIFFERENTIATING VERSUS NO ADVERTISING CONDITIONS 



SOURCE 

ADVERTISING 
SEQ 

AD*SEQ 
SS/AD,SEQ 



TYPE III SS DF 

0.01854948 1 

0.13948513 15 

0.24968350 15 

0.44225911 32 



'JP 


^LUE 


P > F 


1 


34 


0.2552 





67 


0.7907 


1 


20 


0.3180 



104 



DF 


F VALUE 


P > F 


1 


6.68 


0.0145 


15 


0.87 


0.5962 


15 


0.75 


0.7213 


32 







TABLE VT-12 

AVERAGE NUMBER OF BRANDS EVER CHOSEN 
DIFFERENTIATING VERSUS NO ADVERTISING CONDITIONS 



SOURCE TYPE III SS 

ADVERTISING 30.25000000 

SEQ 59.43750000 

AD*SEQ 50.75000000 

SS/AD,SEQ 145.00000000 



ever chosen, because it will lead to some brands being 
perceived to be clearly superior to others. 

Taken together, the pattern of results discussed in this 
section indicate that differentiating advertising did 
significantly affect brand-utilities in the stimulus-based 
decision environment. The increased dispersion in brand- 
utilities can be expected to have implications for price 
elasticities observed under this condition. 

Effects of reminder advertising . A multivariate 

contrast between the reminder advertising condition and the 
no advertising condition in a MANOVA analysis of the four 
preference measures showed no significant differences F(4,29) 
< 1. 

Univariate tests reveal that the effects of reminder 
advertising on the measures of interbrand variance in 
preferences approached significance in some cases. Reminder 
advertising increased the standard deviation of choice 



105 



frequencies (M = 4.93) compared to the no advertising 
condition (M. = 4.25), and this effect approached 
significance, F{1,62) = 2.67, p. < .11, (oo^ = .02). The 
results are presented in Table VI-13. 

The dispersion in the proportion of budget allocated to 
the various brands were lower in the reminder advertising 
condition (M = .09) than in the no advertising condition (M = 
.08), and this difference approached significance F(l,62) = 
3.16, p. < .08, (CO = .02). These results are presented in 
Table Vl-14. 

An analysis of variance on the share of budget allocated 
to the top brand across these two advertising conditions was 
not significant, F(1.62) = 2.12, p < .15, (CO^ = .01). These 

results are presented in Table VI-15. Results (presented in 
Table VI-16) show that the average number of brands ever 
chosen did not significantly differ between the reminder 
advertising condition and the no advertising condition, 
F(l, 62) = 2.22, p < .15. 



TABLE VI-13 

DISPERSION IN CHOICE FREQUENCIES 
REMINDER VERSUS NO ADVERTISING CONDITIONS 



SOURCE TYPE III SS 

ADVERTISING 7.51580254 1 2.67 0.1118 1 



SEQ 39.84124419 

AD*SEQ 29.03591569 

SS/AD,SEQ 89.95440461 



DF 


F VALUE 


P > F 


1 


2.67 


0.1118 


15 


0.94 


0.5259 


15 


0.69 


0.7763 


32 







106 



TABLE VI- 14 

DISPERSION IN PROPORTION OF BUDGET 
REMINDER VERSUS NO ADVERTISING CONDITIONS 



SOURCE 



TYPE III SS 



DF 



F VALUE 



P > F 



ADVERTISING 
SEQ 

AD*SEQ 
SS/AD,SEQ 



0.00274298 1 

0.01727033 15 

0.01191580 15 

0.01191580 32 



3.16 
1.33 
0.92 



0.0848 
0.2429 
0.5561 



TABLE VI- 15 

SHARE OF BUDGET ALLOCATED TO TOP BRAND 
REMINDER VERSUS NO ADVERTISING CONDITIONS 



SOURCE 



TYPE III SS 



DF 



F VALUE 



P > F 



ADVERTISING 





01930218 


1 


SEQ 





15181354 


15 


AD* SEQ 





17488103 


15 


SS/AD,SEQ 





29074490 


32 



2.12 
1.11 
1.28 



0.1547 
0.3835 
0.2681 



TABLE VT-16 

AVERAGE NUMBER OF BRANDS EVER CHOSEN 
REMINDER VERSUS NO ADVERTISING CONDITIONS 



SOURCE 


TYPE III SS 


DF 


F VALUE 


P > F 


ADVERTISING 


11.39062500 


1 


2.22 


0.1464 


SEQ 


66.48437500 


15 


0.86 


0.6081 


AD*SEQ 


38.35937500 


15 


0.50 


0.9240 


SS/AD,SEQ 


164.50000000 


32 







107 



It is interesting to note that although reminder 
advertising was not postulated to significantly influence the 
perceived utilities offered by the various brands, the 
results indicate that it did have a minor effect. Ananalysis 
of debriefing protocols suggest that the picture of the 
package provided some information about the relative worth of 
the various brands, which in the absence of other information 
was used as a means of differentiation. Thus, it is likely 
that the reminder ads used for the study served more than 
just the reminder function. 
Price Elasticity 

Price elasticity was calculated for each of the 12 
brands at the individual subject level by looking at 
quantities purchased across 16 shopping trips. Prices were 
varied over the trips, and the quantities purchased for each 
brand at their respective high and low prices were aggregated 
to compute price elasticity. The arc elasticity measure that 

was used is defined as 

AQ/Mean Q 

Arc Elasticity = 

AP/Mean P 

For each subject, the twelve brand-level elasticities 
were analyzed with brands as a within-sub jects factor. It 
was expected that the 12 brands should show similar patterns 
of these manipulations, although main effects of brands might 
be based upon differential popularity of the brands. 



108 



ANOVA effects treating brands as fixed . The data were 
analyzed in a mixed ANOVA design with between-subject s 
factors of Advertising Condition, Decision Environment and 
Sequence, with Brands as a repeated factor (3 X 2 X 16 X 12) . 
Results are first presented treating all factors, (including 
Brands) as fixed. This would imply that statistical 
generalizations from this experiment are limited to effects 
observed with the specific brands that were used in the 
study. The results are presented in Table VI-17 to VI-19. 

Results indicate that there was a significant effect of 
decision environment, advertising condition and the 
Advertising X Environment interaction. Analyses of planned 
contrasts were conducted to better understand the locus of 
these effects on price elasticity. 

TABLE VI-17 
TESTS OF HYPOTHESES FOR BETWEEN SUBJECTS EFFECTS 

SOURCE TYPE III SS DF 

ENVIRONMENT 190.80058080 

ADVERTISING 85.14927157 

ENV*AD 252.53599766 

SEQ 92.52268851 

ENV*SEQ 136.42643032 

AD*SEQ 248.41869797 

ENV*AD*SEQ 310.57843835 

SS/ENV,AD,SEQ 820.53705672 95 



DF 


F VALUE 


P > F 


1 


22.09 


0.0001 


2 


4.93 


0.0092 


2 


14.62 


0.0001 


15 


0.71 


0.7646 


15 


1.05 


0.4102 


30 


0.96 


0.5357 


30 


1.20 


0.2515 



109 



TABLE VI- 18 

ANALYSIS OF VARIANCE 
UNIVARIATE TESTS OF HYPOTHESES FOR WITHIN SUBJECTS EFFECTS 



SOURCE 




TYPE III SS 


DF 


F VALUE 


P 


> F 


BRAND 




152 


29064736 


11 


4 


03 





.0001 


BRAND *ENV 




23 


66011338 


11 





63 





.8075 


BRAND* AD 




56 


13321359 


22 





74 





7971 


BRAND *ENV* 


AD 


124 


03344868 


22 


1 


64 





0316 


BR*SEQ 




504 


17150129 


165 





89 





8275 


BR*ENV*SEC 


) 


615 


79173623 


165 


1 


09 





2309 


BR*AD*SEQ 




1274 


17431349 


330 


1 


12 





0905 


BR*ENV*AD* 


SEQ 


950 


01510608 


330 





84 





9732 


BR*SS/ENV, 


AD,S 


2 3589 


.46828 


1045 











TABLE VI-1 9 
MEAN PRICE ELASTICITY 



ENVIRONMENT 



ADVERTISING 
NONE 

REMINDER AD 
DIFFERENTIATING 



STIMULUS 
BASED 



■2.02 
■1.54 
■1.24 



MEMORY 
BASED 



-0.53 
-1.59 
-0.94 



In the absence of advertising, price elasticity was 
greater in the stimulus-based environment (M = -2.02) than in 
the memory-based environment (M = -0.53), F(l,95) =36.44, ^ < 
.001, (CO = .08) . This finding lends strong support to HI 

that the price elasticity will be greater under stimulus- 



110 



based choice situations than under memory-based choice 
situations . 

Price elasticity in st imulu s -based decision 
environments . Analyses of planned contrasts were carried out 
to test for the effects of advertising on price elasticity in 
the stimulus-based choice environment. Results (presented 
in Tables VI-20 and VI-21) show that there was a significant 
simple effect of advertising on price elasticity, F(2,95) = 
6.90, £ < .01,(00 = .04) lending support to H2 . Because 

there were no significant differences in the average size of 
the consideration set across the three advertising 
conditions, it is possible to conclude that these effects on 
price elasticity in the stimulus-based environment are 
mediated by changes in the variance in brand-utilities as a 
result of advertising. 

The difference in price elasticity differentiating 
advertising condition (M = -1.24) and the control condition 
(M = -2.02) was significant, F(l,95) = 13.53, t^. < .001. This 
finding supports H2a that differentiating advertising will 
result in lower price elasticity. This parallels the results 
obtained for preferences, where it was found that advertising 
that seeks to differentiate brands resulted in increased 
variance in the perceived utilities of the brands. 

H2a predicted that there will be no difference in price 
elasticity between the reminder advertising condition and the 
control condition. However, results showed that price 



Ill 



elasticity was significantly lower in the reminder 
advertising condition (M = 1.54) compared to the control 
condition (M = 2.02), F(l,95) = 5.28, u < -025. Thus there 
was no support for H2b. These results are consistent with 
the findings that reminder advertising did have some effect 
on the variance in utilities in the stimulus-based 
environment, discussed in the previous section. It is 
possible that there was an element of differentiation even in 
the reminder ads. An analysis of debriefing protocols 
suggest that they did provide some information about the 
brand, which, in the absence of any other information became 
a basis for elimination and choice. Moreover, to the extent 
package information cued the recall of information learned 
during the first session, it is perhaps expected that 
reminder ads would lead to some changes in preferences. 

Results presented in Table VI-22 showed that price 
elasticity was marginally less in the differentiating 
advertising condition (M = 1.24) compared to the reminder 
condition, (M= 1,54), F(l,95) = 1.90, ;^< .15. 

In the stimulus-based condition, there were no 
significant effects of advertising on the size of the 
consideration set. Thus, when the brand names were readily 
available, the presence or absence of advertising or the 
information provided by advertising did not have an effect on 
the number of brands searched. However, there were 
significant effects of advertising on preferences. Since 



112 



the effects on price elasticity are very similar to the 
pattern of results obtained for preferences, it is possible 
to conclude that, in stimulus-based environments, the 
differential preference for the various brands is a key 
mediator of price elasticity. This result provides strong 
support for H2 , that the only effect of advertising in the 
stimulus-based choice situation will be mediated by the 
dispersion in brand-utilities. 

TABLE VI-2Q 

PRICE ELASTICITY IN STIMULUS-BASED ENVIRONMENT 
DIFFERENTIATING VERSUS NO ADVERTISING CONDITIONS 

SOURCE TYPE III SS 

ADVERTISING 116.86385898 
SEQ 79.59978460 

AD*SEQ 238.97790555 
SS/ENV,AD,SEQ 820.53705672 



TABLE VI-21 

PRICE ELASTICITY IN STIMULUS-BASED ENVIRONMENT 
REMINDER VERSUS DIFFERENTIATING ADVERTISING CONDITIONS 



SOURCE TYPE III SS 

ADVERTISING 45.52724071 
SEQ 122.47040230 

AD*SEQ 89.24442242 

SS/ENV,AD,SEQ 820.53705672 



DF 


F VALUE 


P>F 


1 


13.53 


0.001 


15 


0.61 


0.904 


15 


1.84 


0.133 


95 







DF 


F VALUE 


P>F 


1 


5.28 


0.025 


15 


0.72 


0.750 


15 


0.52 


0.909 


95 







113 



TABLE VI-22 

j PRICE ELASTICITY IN STIMULUS-BASED ENVIRONMENT 

REMINDER VERSUS NO ADVERTISING CONDITIONS 



SOURCE TYPE III SS DP 



DP 


F VALUE 


P>F 


1 


1.90 


0.155 


15 


0.80 


0.788 


15 


0.77 


0.769 


95 







ADVERTISING 16.45361206 
SEQ 103.76050235 

AD*SEQ 100.72778704 
SS/ENV,AD,SEQ 820.53705672 



Price el asticity in memory-based decision environments . 
Under the memory-based condition there were significant 
differences in price elasticity across advertising 
conditions, F(2,95) = 12.72, ^< .001, (ro^ = .04). Thus the 

evidence supports H3, that there will be a significant simple 
effect of advertising on price elasticity in the memory-based 
condition. Tables VI-23 to VI-25 present the results of the 
analyses of planned comparisons to test for the effects of 
reminder and differentiating advertising on price 
sensitivity. 

Compared to the no advertising condition, price 
elasticity was greater in the differentiating advertising 
condition (M = -0.94), but this difference was not 
significant, F(l,95) = 3.70, p. < .10. Thus there was no 
support for H3a that differentiating advertising would reduce 
price elasticity. 



114 



1 


3.70 


0.102 


15 


1.56 


0.144 


15 


1.02 


0.463 


95 







TABLE VI-23 

PRICE ELASTICITY IN MEMORY-BASED ENVIRONMENT 
DIFFERENTIATING VERSUS NO ADVERTISING CONDITIONS 



SOURCE TYPE III SS DF F VALUE P>F 

ADVERTISING 31.96538463 
SEQ 131.04664046 

AD*SEQ 85.60967106 

SS/ENV,AD,SEQ 820.53705672 



TABLE VI-2 4 

PRICE ELASTICITY IN MEMORY-BASED ENVIRONMENT 
REMINDER VERSUS NO ADVERTISING CONDITIONS 



SOURCE TYPE III SS 

ADVERTISING 216.49080913 
SEQ 118.40784879 

AD*SEQ 130.87140122 
SS/ENV,AD,SEQ 820.53705672 



DF 


F VALUE 


P>F 


1 


25.07 


0.001 


15 


0.91 


0.797 


15 


1.23 


0.251 


95 







TABLE VI-25 

PRICE ELASTICITY IN MEMORY-BASED ENVIRONMENT 
REMINDER VERSUS DIFFERENTIATING ADVERTISING CONDITIONS 



SOURCE 


TYPE III SS 


DF 


F VALUE 


P>F 


ADVERTISING 


78.06047822 


1 


9.03 


0.015 


SEQ 


184.48337550 


15 


1.42 


0.158 


AD*SEQ 


210.02896638 


15 


1.62 


0.109 



SS/ENV,AD,SEQ 820.53705672 95 



115 



The findings show that price elasticity was greater 
under reminder advertising (M = -1.59) compared to the 
control condition (M = -0.53), F(l,95) = 25.07, p. < .001. 
Thus result supports H3b which predicted that reminder 
advertising will increase price elasticity. This parallels 
the the effects of reminder advertising on the size of the 
consideration set. 

Results presented in Table VI-25 show that price 
elasticity was significantly greater in the reminder 
advertising condition (M = 1.59) compared to the 
differentiating advertising condition (M = 0.94), F(l,95) = 
1.03, T^ < .01. 

ANOVA effects treating brands as random . Thus far the 
analysis has treated brands as a fixed factor. In such 
cases, statistical generalizations are limited to the 
treatment effects observed with the specific treatment 
conditions used in the experiment. In order for statistical 
conclusions to extend to brands in general, the twelve level 
brands factor has to be treated as random. This implies that 
the structural model appropriate for this design is different 
from the one that was used for the case where brands were 
treated as fixed. 

When the brands factor is treated as random, the usual 
error terms cannot be used to estimate the between-sub jects 
effects of interest to the study (Keppel 1982) . Therefore 
the denominator in the F-ratio is a combination of terms 



116 



which together provide an appropriate test of the treatment 
effect. The terms which were combined to create the 
appropriate errors terms for the hypothesized effects are 
presented in Appendix HE. These quasi F ratios have degrees 
of freedom that are adjusted according to Satterthwaite ' s 
(1946) method. 

The results are similar to those obtained when brands 
were treated as fixed. There was a main effect of 
advertising condition, a main effect of decision environment 
and an environment X advertising condition interaction. The 
results are presented in Table VI-26. 

TABLE VI-2 6 

ANALYSIS OF VARIANCE 

TESTS OF HYPOTHESES FOR BETWEEN SUBJECTS EFFECTS 

BRANDS AS RANDOM FACTOR 

SOURCE SS DF DFe F P > F 

ENVIRONMENT 190.8005 1 45 25.92 0.001 

ADVERTISING 85.1492 2 55 5.48 0.009 

ENV*AD 252.5359 2 52 11.64 0.001 

SEQ 92.5226 15 80 0.75 0.764 

ENV*SEQ 136.4264 15 91 1.02 0.412 

AD*SEQ 248.4186 30 98 0.91 0.536 

ENV*AD*SEQ 310.5784 30 79 1.28 0.251 



Note : DFe implies the adjusted error degrees of freedom used 
for the analysis. 



117 



An analysis of planned comparisons revealed that there 
was a simple main effect of advertising in the stimulus-based 
choice environment, F.(2,55) = 7.69, p. < .01. There were 
significant differences in price elasticity when subjects 
were exposed to differentiating advertising, compared to the 
situation when they were not exposed to any ads, F(l,55) = 
15.05, 2. < .001. This result supports H2a. There were also 
significant differences in price elasticity between the 
reminder ad and the control condition, F(l,55) = 5.88, p. 
<.025. The results were in support of HI that advertising 
will have a significant effect on price sensitivity in the 
stimulus-based environment. The difference in price 
elasticity between the reminder advertising condition and the 
differentiating advertising condition was marginally 
significant, F(l,55) = 2.20, ^ < .10. 

In the memory-based choice environment, there was a 
simple main effect of advertising on price elasticity, 
11(2,55) = 14.16, p. < .001. There were significant 
differences between reminder advertising and differentiating 
advertising in their respective effects on price elasticity. 
There was a significant positive effect of reminder 
advertising, F(l,55) =27.9, £ < .001. This result supports 
H3a. Also, differentiating advertising had a positive effect 
on price elasticity compared to the control condition, 
F(l,55) = 4.12, £ < .05. Price elasticity was significantly 
greater in the reminder advertising condition compared to the 



118 



differentiating advertising condition, F(l,55) = 10.06, p. < 
.001. 

In the absence of advertising, there was greater price 
sensitivity in the stimulus-based decision environment than 
in the memory-based environment, F(l,45) = 55.16, £ < .001. 

To conclude, when brands are treated as a random factor 
there is no major change observed in the pattern of results. 
The major hypotheses of the experiment were supported and 
therefore it was possible to demonstrate that advertising 
does affect price sensitivity. The nature of the effect of 
advertising depends on nature of the advertising and also the 
information that is available at the time of choice. It is 
possible to assert that interactions of the major results 
with brands are sufficiently small that the results obtained 
need not be limited to the brands used for the study, but are 
likely to hold in a more general case. 

Chapter VII presents a general discussion of the results 
obtained in Experiment 3. Implications of the results for 
research on the economic effects of advertising are also 
discussed. 



CHAPTER VII 
GENERAL DISCUSSION AND IMPLICATIONS 

This dissertation investigated the relationship between 
advertising and consumer price sensitivity. This topic has 
been central to the controversy about the economic effect of 
advertising in the economics literature (Comanor and Wilson 
1979, Ornstein 1977) . Two competing theories in economics 
have made conflicting predictions about the potential effect 
of advertising, based on assumptions about consumer responses 
to this marketing mix variable. The fragmentary empirical 
evidence in economics and marketing literature has not 
provided conclusive support for either theory. Previous 
researchers have tested deductions from these theories and 
have not been able to offer any test of the underlying 
theoretical assumptions. Thus there remain controversies 
over whether advertising increases or decreases price 
elasticity based on assumptions about its informative or 
persuasive nature, with conclusions drawn from studies that 
were not designed to provide answers about causality. 

This dissertation proposed that prior conceptualizations 
about the effect of advertising on consumer price sensitivity 
are simplistic. It has been suggested in the literature that 
the diverging empirical studies on this topic have not taken 
account of the content of advertising (Lambin 1976, Farris 



119 



120 



and Albion 1980) . Yet research in this area has not 

explicitly examined the effects of variations in advertising 

themes on consumer behavior. This study postulated that it 

is the content and not simply the volume of advertising as 

has been traditionally studied, that determines consumer |^ 

responses . [ 

It has been suggested (Lynch & Bloom 1987) that the [; 

effect of advertising on consumer price sensitivity can be 

better understood by an examination of the role of 

advertising in the choice decision. This dissertation has 

\ 
empirically examined the framework offered by these authors 

and investigated the basic causal mechanisms underlying the 

effect of advertising on consumer price sensitivity in an | 

( 
attempt to understand and predict the direction of change. 

Two key theoretical constructs were postulated to mediate the 

relationship. The first is the size of the consideration 

set, which consists of brands that the consumer actively 

considers at the time of choice. The second is the 

distribution of preferences for brands in the consideration 

set . 

It was proposed that different types of advertising 

affect both constructs, but have potentially different 

effects on the relative importance of these two mediating 

variables. For example, differentiating advertising has 

stronger effects than reminder advertising on the 

distribution of brand-utilities, but reminder advertising has 



121 



larger effects than differentiating advertising on the number 

of alternatives considered at the time of choice. It was 

suggested that the effect of advertising on the size of the 

consideration set is moderated by whether the decision 

environment is stimulus-based or memory-based. That is, the 

more the consumer relies upon memory to generate a set of 

candidate brands, the more important will be the tendency for 

advertising to increase the size of the consideration set by 

facilitating recall. This will lead to an increase in price 

elasticity. Insofar as the consumer can and does rely on 

external information in the decision environment to generate 

a consideration set, the facilitating effects of advertising 

will be mitigated. In such a case, advertising's primary 

effect will be to make demand more inelastic by producing 

larger perceived differences in utility among the various 

brands. Thus the primary contribution of the dissertation is 

to document the effects of advertising on these two mediating 

constructs that were hypothesized to have direct implications 

for the nature of the relationship between advertising and 

price elasticity. 

Effects of Ad vertising on Distribution of Brand-Utilities 

The study examined the effect of advertising on the 
variance of interbrand utilities. It was expected that 
advertising that resulted in real or ' image ' -based 
differentiation would increase the dispersion in utilities, 
leading to some brands being perceived to be superior to 



122 



others. Results from the analyses of the four measures of 
interbrand variance of revealed preference suggest that 
differentiating advertising did have a significant influence 
on the dispersion of brand-utilities in the stimulus-based 
decision environment. This increased dispersion in brand- 
utilities can be expected to have implications for price 
elasticities observed under this condition. 

It is interesting to note that although reminder 
advertising was not expected to significantly affect the 
perceived utilities offered by the various brands, the 
results indicate that it did have a minor effect. It is 
possible that the picture of the package and the brand names 
were used as a basis for preference formation. Also, such 
ads could have cued the recall of information about 
attributes learned in the first session. Keller (1987) shows 
that such recall cues have an effect on the spread of memory- 
based evaluations. This suggests that even 'reminder' 
advertising has the potential for affecting the distribution 
of perceived interbrand utilities. 

In conclusion, all advertising has a reminder function, 
that is, to provide recall cues in order to ensure that the 
brand is retrieved at the time of choice. Advertising also 
aims at differentiating the brand from competitive offerings 
so as to ensure that the brand is preferred to others that 
the consumer might consider. With respect to the first 
objective, both reminder and differentiating advertising had 



123 



f 



a positive effect on the recall of brand names. With respect i 
to the second objective, the results suggest that 
advertising-induced differentiation is a matter of degree. 
Hence, the more the utility connoting information in an ad, 
the more is its differentiating function over and above its 
reminder function. Therefore in this experiment, 
differentiating advertising led to an increased variance of 
utilities, compared to reminder ads. These findings are 
likely to have implications for price elasticity. 
Effects of Advertising on the Size of the Consideration Set 

It was proposed that the effects of advertising on price 
elasticity could be better understood by examining its 
effects on the size of the consideration set. The pattern of 
results provide interesting insights on how advertising might 
affect the size of the consideration set by influencing the 
ability and the motivation to retrieve brand names for 
consideration. The findings suggest that the effects of 
advertising on the consideration set could differ, depending 
on the type of advertising and the decision environment. 

Advertising effects on knowledge and remembrance of 

substitutes versus perceived substitutability of brands . 

Price elasticity is a function of perceived substitutability 
(Nelson 1974a, 1978) . Economists of the information school 
have assumed that advertising provides information about 
brands and increases the number of known substitutes. This 
implies that if advertising has made consumers 'aware' of a 



124 



particular brand when it was first introduced, this should be 
sufficient for it to be considered at the time of choice. 
Consumer researchers (Alba and Chattyopadhyay 1985; Nedungadi 
1987) have suggested that mere awareness of a brand does not 
necessarily ensure that the brand is retrieved at the time of 
choice - the brand has to be retrieved from long-term memory. 
It is unlikely that all brands in a product category are 
effective substitutes in a particular choice decision. 
Consumers faced with an assortment of brands attempt to focus 
on a subset of alternatives in order to gain efficiency in 
shopping. As this experiment has demonstrated, mere 
awareness of a brand is not sufficient for it to be included 
in the consideration set. Advertising was shown to increase 
the number of effective substitutes by increasing the 
knowledge and remembrance of brand alternatives. 

Advertising that emphasizes real or image-based 
differentiating attributes provides information about the 
product on the basis of which overall evaluations are made. 
In this experiment, it was shown that such advertising makes 
some brands perceived to be superior to others and therefore 
reduces perceived subst itutability among brands. This 
implies that the number of effective substitutes from which 
consumers makes a choice is reduced, and therefore a larger 
price discount will be required in order to get them to 
switch from their most preferred brands. 



125 



Memory for bran ds and the size of the consideration set . 
The results from Experiment 1 and 2 show that in memory-based 
decision environments, both reminder and differentiating 
advertising enhance the ability to recall brand names. 
Experiment 3 shows that this has a positive effect on the 
size of the consideration set. The significant differences 
in consideration set size in the advertising conditions as 
compared to the control condition lend strong support to the 
hypothesis that advertising makes brand names more accessible 
in memory, and thus increases the size of the consideration 
set. This memory-enhancing capability of advertising holds 
for differentiating advertising as well. 

Dispersion of brand-utilities and the size of the 

consideration set. Besides affecting the memory of brand 

names, advertising also increased the variance in interbrand 
preferences. It has been noted earlier that differentiating 
ads contained more utility connoting information than 
reminder ads and thus resulted in more significant increases 
in the dispersion of perceived utilities. This is reflected 
in the size of the consideration set in the memory-based 
condition . 

Recall that the results of Experiment 2 showed that 
there are no significant differences between reminder and 
differentiating advertising on the ability to recall brand 
names. They were equivalent in that they both increased the 
ability to recall brand names from memory by increasing the 



126 



accessibility of these names. If consideration set size is a 
function of ability alone and is unaffected by the motivation 
to recall brand names from memory, then there should be no 
differences between the two advertising conditions in the 
size of the consideration set. However, results from 
Experiment 3 showed that consideration sets were 
significantly smaller in size under the differentiating 
advertising condition compared to the reminder advertising 
condition. This suggests that the information contained in 
the two types of advertising differentially affected the 
motivation to retrieve brand names in a prepurchase situation 
under which the objective was to search for prices. In 
short, different cues or criteria are used for consideration. 

When brand names have to be generated form memory, the 
most preferred brands are likely to be recalled first 
(Nedungadi and Hutchinson 1985) . Under such a scenario the 
expected marginal benefit of adding brands for consideration 
are expected to be low, because is is unlikely that a price 
discount on a less favored brand will be able to compensate 
for the perceived superiority of the top brands. Results 
showed that differentiating advertising resulted in higher 
perceived variance of interbrand preferences compared to 
reminder advertising. Consistent with this finding, when 
brand names had to be generated from memory, consideration 
set sizes were significantly smaller when advertising was of 
the differentiating variety. 



127 



Costs of thinking . Even though differentiating and 

reminder advertising differentially affected the interbrand 
dispersion of utilities, this was not reflected in 
significant differences in the size of the consideration set 
in the stimulus-based situation. This is in contrast to the 
memory-based case discussed in the preceding section. There 
is no evidence to support the prediction that differentiating 
advertising will reduce the size of the consideration set 
more than reminder advertising in the stimulus-based decision 
environment . 

Because subjects had a list of brand names available to 
them, it is possible that the perceived marginal cost of 
thinking (Hauser & Wernerfelt 1990; Shugan 1980) associated 
with adding another brand to the consideration set was 
relatively low in stimulus-based decisions. Therefore it 
would require more advertising induced dispersion of brand- 
utilities to discourage subjects from at least considering 
more brands, even though some brands were more preferred and 
thus had a greater probability of being chosen. This would 
explain why, under such conditions, differentiating 
advertising did not result in shrinking the size of the 
consideration set, even though it affected the variance of 
brand-utilities . 



128 



Combined Effects of Consideration Set Size and Brand-Util J t.y 
Dispersion on Price Elasticity 

The effects of advertising on price elasticity was 
postulated to be mediated by two key constructs: the size of 
the consideration set and the variance of interbrand 
preference. The results on these theoretical constructs 
discussed in the preceding sections enable a better 
understanding and interpretation of advertising-induced 
changes in price elasticity. 

In the memory-based decision environment, both reminder 
and differentiating advertising resulted in an increase in 
price elasticity. This is consistent with the finding that 
advertising (both reminder and differentiating) increases the 
size of the consideration set in memory-based environments. 
When brand names have to be retrieved from memory, 
advertising (both reminder and differentiating) increases the 
number of brands considered by making brand names more 
accessible in memory. Given the assumption that price 
elasticity is a positive function of consideration set size, 
this implies that price elasticity might increase even with 
differentiating advertising. 

It was noted earlier that differentiating advertising 
resulted in smaller consideration sets compared to reminder 
advertising. This finding is consistent with the result that 
differentiating advertising led to a greater variance in 
interbrand preferences. The differences in price elasticity 



129 



between the two advertising conditions reflect the indirect 
effects of the dispersion of utilities on the price 
elasticity of demand. However, the recall-facilitating 
effect of differentiating advertising was more powerful, so 
that there was an overall increase in the size of the 
consideration set even under differentiating advertising. 
Thus, compared to the no advertising condition, 
differentiating advertising led to greater price sensitivity. 

In the stimulus-based condition, there were no 
significant effects of advertising on the size of the 
consideration set. The presence or absence of advertising or 
the information provided by advertising did not have an 
effect on the number of brands searched when the brand names 
were readily available. On average, consumers searched for 
information on about roughly one-third of the available 
brands. As has been noted earlier, in this situation the 
facilitating effects of advertising on recall are not as 
important as in the memory-based case. 

However, there were significant effects of advertising 
on the dispersion of brand-utilities. Since the effects on 
price elasticity are very similar to the pattern of results 
obtained for the dispersion of preferences, it is possible to 
conclude that in stimulus-based environments, the 
differential preference for the various brands is a key 
mediator of price elasticity. 



130 



The significant differences in price elasticity between 
the differentiating advertising condition and the control 
condition strongly support the prediction that 
differentiating advertising increases the interbrand 
variance in preferences and thereby decreases price 
elasticity. As compared to the control condition, even 
reminder advertising reduced price elasticity in the 
stimulus-based environment. This is consistent with the 
finding that reminder advertising influenced the dispersion 
of utilities, although the effect was not as large as in the 
case of differentiating advertising. The results as a whole 
suggest that when brand names are readily available and 
consumers rely on this information to generate the 
consideration set, the effects of advertising on price 
elasticity are primarily mediated by its effect on the 
distribution of perceived utilities for the various brands. 

From these results, it is possible to conclude that when 
brand names have to be recalled from memory, price elasticity 
is a function of both the size of the consideration set and 
the dispersion of brand-utilities. Advertising has a 
potential effect on both these constructs. The net impact of 
advertising on price elasticity therefore depends on the 
strength of its effects on these two variables. In the 
stimulus-based environment, the effect of advertising on 
price elasticity is a function of its effect on interbrand 
variance in preference. 



131 



Implications 

This research has demonstrated that it is possible to 
make predictions about the potential effects of advertising 
on price elasticity by examining its effects on the two key 
constructs. They are the size of the consideration set and 
the dispersion of brand-utilities. Price elasticity is 
positively related to the size of the consideration set and 
negatively related to the dispersion of utilities across the 
various brands. These two variables mediate the relationship 
between advertising and price elasticity, and the net effect 
depends on the strength of the parameters. Different kinds 
of advertising and the decision environment have different 
effects on these parameters. 

Environmental factors affecting the mediating role nf 

consideration set size . In cases in which brand name recall 
is not important for the consumer to generate the 
consideration set, the effects of advertising on the memory 
for brand names is not likely to be the major determinant of 
the size of the consideration set. For example, if a 
customer relies on the salesperson's recommendations for 
purchasing a durable good, then the reminder function of 
advertising has a lesser role. Here, advertising's effects 
on price sensitivity will be primarily based on ad-induced 
dispersion of preferences. The greater the variance in 
preference for the brands caused by advertising, the lesser 
will be the price sensitivity. 



132 



The more that the consideration set has to be generated 
from memory, the greater will be the reminder effects of 
advertising on the size of the consideration set. The larger 
the effects of advertising on the consideration set size, the 
greater is the price elasticity for the other brands in the 
consideration set. For example, consumer try to recall the 
stores at which they could shop, and all such stores are not 
arrayed in front of them, they have to rely on memory to 
retrieve names of the stores. To the extent that some stores 
advertise, they are likely to be recalled and as a result get 
included in the consideration set. In such a case, 
advertising performs an important reminder function and its 
effect on price elasticity is greatly influenced by its 
effect on the size of the consideration set. 

Environmental factors mediating the role of dispersion 
of — brand-utilities . In this research, the brands were 
unfamiliar to the subjects. Therefore it was relatively easy 
: to change the distribution of preferences through 

advertising. But in situations in which consumers have prior 
familiarity with the product, the effects of advertising on 
the dispersion of utilities is likely to be smaller. 
Therefore advertising induced differentiation will have a 
lesser effect on price elasticity. In such a case, if the 
recall-facilitating aspect of advertising is important in 
determining the size of the consideration set, then even 



133 



differentiating advertising might lead to an increase in 
price elasticity. 

This research suggests that the mechanism by which 
advertising effects on price sensitivity when all brand names 
are externally available might be quite different from that 
when the names are retrieved from memory. Results showed 
that advertising influences the ability to retrieve brands in 
the memory-based situation, with much smaller effect on the 
stimulus-based situation. It also differentially affects the 
motivation to consider more brands. Because the cost of 
thinking associated with adding another brand to the 
consideration set may be less in stimulus-based decisions, it 
requires more advertising-induced dispersion of brand- 
utilities to discourage consumers from at least considering 
brands . 

Moreover, in the memory-based environments, more 
preferred brands are likely to be recalled first. Thus if 
advertising increases the dispersion of interbrand 
preferences, it will reduce the size of the consideration 
set. To the extent that in stimulus-based environments, 
order of search is guided by the organization of the display 
(Bettman & Kakkar 1977; Biehal & Chakravarti 1982; Alba et 
al. 1990) such that the most preferred brands are not 
considered first, increases in interbrand variance resulting 
from advertising should increase the motivation to consider 



134 



more brands (Nelson 1970, 1974a, 1975; Hauser & Wernerfelt 
1990) . 

The stimulus-based choice situation was used as the 
baseline to demonstrate that consideration sets are likely to 
be situation-specific. Here, he same differentiating ads did 
not lead to any significant reduction in the size of the 
consideration set, even though they had significant effects 
on the dispersion of brand-utilities. It is likely that in 
the laboratory setting, subjects were more motivated than the 
average consumer. In this situation in which perceived costs 
were less than expected benefits, the consideration set size 
was not significantly reduced. Therefore the effects of 
advertising on price elasticity were mediated by the 
dispersion of brand-utilities. 

It can be argued that in real-world situations the 
information processing costs of exhaustively considering and 
evaluating all brands for choice are much higher than in the 
laboratory setting. Therefore, even decisions which are in 
principle, stimulus-based have a large memory component (Alba 
et al . 1990). Consumers need to concentrate on a subset of 
alternatives in order to gain efficiency in shopping. For 
example, the consumer making purchases in the branded 
packaged goods category does not have the time nor the 
motivation to scan every item on the supermarket shelf before 
making a decision (Park et al 1989) . In such a situation, 
the reminder function of advertising becomes crucial in 



135 



getting a brand considered for choice. This will have 
implications for price elasticity. 

The memory effects of advertising are even more dramatic 
in the case of situations which are necessarily memory-based, 
as in the case of the store choice decision discussed 
earlier. In this case, advertising can ensure that a brand 
is included in the consideration set and thus have a 
significant effect on price elasticity. 
Managerial Implications 

The dissertation highlights certain important facts 
about the effect of advertising on consumer behavior which 
could be useful to managers. It provides insights on the two 
major functions of advertising, which combine to produce its 
effects on price sensitivity. Therefore, the difference 
between reminder and differentiating advertising is more a 
matter of degree. Depending on the environment in which the 
decision is made, advertising effects memory for brand names 
and/or preferences - and the relative strength of these two 
parameters determines its net effect on price elasticity. 
This suggests that a differentiating strategy need not 
necessarily result in reduced price sensitivity. As this 
study showed, if the effects of advertising on memory are 
substantial, then the increase in price elasticity as a 
result of the larger consideration set size can offset the 
reduction due to the advertising-induced variance in 
preferences . 



136 



Implications for Public Policy 

This study has considered the economic effect of 
different types of advertising strategies which are likely to 
be adopted by marketers. The results of the study have 
implications for public policy. It suggests that it is 
difficult to theorize about the economic effect of 
advertising in terms of general statements about whether 
advertising increases or decreases price elasticity of 
demand. 

The study supports the view of the information school 
that advertising does have some economic value; it provides 
information on the existence of substitutes. In addition, it 
demonstrates that advertising provides recall cues and 
thereby increases the number of effective substitutes that 
are considered at the time of choice. Price elasticity is a 
function of perceived substitutability, therefore advertising 
by increasing the number of brands considered, can increase 
price elasticity. 

? The market power view of advertising postulates 

advertising to reduce price sensitivity by artificially 
dxfferentiating brands. This study demonstrated that even 

, when advertisers pursue a differentiating strategy, the 

I 

resulting effects on price sensitivity need not always be 

negative. The net effects of advertising on price elasticity 

depend on the strength its effect on the two mediator 

variables postulated in this dissertation. 



137 



Future Research 

Effects of prior experience . This study considered the 
effects of advertising on price elasticity for brands which 
are unfamiliar. It is likely that, in the case of familiar 
brands, the effects of advertising-induced dispersion of 
interbrand utilities on price elasticity will be smaller. 
Future research can examine how prior knowledge moderates the 
effects of advertising on the two key constructs. 

Framing effect s on the weights of price and non-price 
attributes . . This study has considered how advertising might 
influence price elasticity by altering the distribution of 
interbrand preference. Another way in which advertising 
might potentially reduce price sensitivity is by increasing 
the salience of non-price attributes and thereby decreasing 
the importance of price. Thus, advertising can cause 
consumers to 'frame' their choices in terms of the evaluative 
criteria suggested by advertisers (Bettman & Sujan 1987; Alba 
& Hutchinson 1990) . It has been reported that increasing the 
salience of product attributes suppresses the ability to 
recall unmentioned attributes (Alba & Chattyopadhyay 1985) . 
Also, by increasing the salience of a particular attribute, 
advertising might ensure that this attribute is used in 
subsequent evaluations, and thereby increase the effective 
weight given to that attribute (Feldman & Lynch 1988) . Taken 
together, this suggests that advertising, by increasing the 
salience of non-price attributes, could lead to decreased 



138 



price sensitivity. Conversely, price advertising, typical of 
supermarket advertising for packaged goods, can increase the 
salience of price and suppress the salience of non-price 
attributes. Bettman and Sujan's (1987) work suggests that 
these effects will be stronger for consumers low in prior 
knowledge. Future research could study the effect of 
advertising price versus non-price attributes on price 
elasticity . 

Effects of interbrand variance in the amount of 

advertising . In this experiment, all brands were advertised, 
and therefore advertising had a significant positive effect 
on recall and consideration set size. This led to an 
increase in price elasticity. In the real world, some brands 
are advertised, and marketers' tactics can increase recall 
deficits of consumers (Alba & Hutchinson 1990) . Research has 
shown that presenting a subset of brands is likely to inhibit 
the recall of competing brands (Alba & Chattyapadhyay 1985) . 
Thus if some big advertisers get a large share of exposures, 
unadvertised brands have a lower chance of being considered 
and therefore of being chosen. Future research can examine 
the effect of unequal levels of advertising by competitors on 
the resulting price elasticity. This would be of interest to 
managers and public policy makers. 



APPENDIX I 



DETAILED LISTING OF STIMULI AND TASKS FOR PRETESTS 



APPENDIX IA-1 
INTRODUCTION AND PRELIMINARY INSTRUCTIONS 

We appreciate your participation in this study which is 
concerned with consumer responses to new brands of candy. 

During this experiment, you will be required to use the 
keyboard to type your responses when prompted by messages on 
the screen. After you have finished typing in your response, 
press the <RETURN> key (marked with the orange sticker) , so 
that your responses may be "received" by the computer. In 
case of a typing error, you may use the <BACKSPACE> key to 
erase the typed letters and retype the correct letters. This 
key is located to the right of the number keys at the top row 
of the keyboard (just above the <RETURN> key) . 

Press <RETURN> to continue 



140 



141 



APPENDIX IA-1 — continued 

This experiment is concerned with consumer responses to 
several new brands of chocolate candy bars which are now 
available in Canada and which are likely to be introduced 
into the Gainesville market . 

In the next few minutes you will be given some 
information on these candy bars. Please pay attention to the 
information. After this, you will be asked some questions on 
them. If you do not answer the questions correctly, you will 
be shown the information once more. 

Press <RETURN> to continue 



142 



APPENDIX IA-1 — continued 

The Company is attempting to assess consumer responses 
to the new brands of chocolate candy bars before the test 
market. Some information on these candy bars, namely, the 
brand name, the weight and what it contains, will now appear 
on the screen. Please pay attention to the information. 

Press <RETURN> to continue.... 



143 



APPENDIX IA-2 
DETAILED LISTING OF BRAND DESCRIPTIONS 



1) 



2) 



BRAND NAME 
WEIGHT: 
CONTAINS : 



BRAND NAME 
WEIGHT: 
CONTAINS : 



Coffee Crisp 

1.9 oz. 

Crisp cream-filled 

biscuit 



Dairy Milk 

1.45 oz. 

Pure milk chocolate 



3) 



4) 



BRAND NAME 
WEIGHT: 
CONTAINS : 



BRAND NAME 
WEIGHT: 
CONTAINS : 



Sweet Marie 

2.1 oz. 

Chewy fudge, caramel and 

peanuts 



Crispy Crunch 

1.75 oz. 

Lots of peanuts in milk 

chocolate 



144 



APPENDIX IA-2 -- continued 



5) 



6) 



7) 



BRAND NAME 
WEIGHT: 
CONTAINS : 



BRAND NAME 
WEIGHT: 
CONTAINS : 



BRAND NAME 
WEIGHT: 
CONTAINS : 



BRAND NAME 
WEIGHT: 
CONTAINS : 



Caramllk 

2.0 oz. 

Bigger bites of soft 

caramel 



Gold 

1.6 oz . 

Crisps and a kiss of 

honey 



Eat-more 

1.4 oz . 

Peanuts in chocolatey 

coating 



Mr. Big 

2.2 oz. 

Wafers, caramel and rice 

crisps 



145 



APPENDIX IA-2 



continued 



10) 



BRAND NAME 
WEIGHT: 
CONTAINS : 



BRAND NAME 
WEIGHT: 
CONTAINS : 



Crunchie 

1.9 oz. 

Golden honeycomb center 



Smart ies 

2.1 oz . 

Candy-coated chocolate 

centers 



11) 



BRAND NAME 
WEIGHT: 
CONTAINS : 



Breakaway 

1.6 oz . 

Crispy wafer and chewy 

caramel 



12; 



BRAND NAME 
WEIGHT: 
CONTAINS : 



Mirage 

1.5 oz . 

Milk chocolate 

yet light 



- thick. 



146 



APPENDIX IA-3 
INSTRUCTIONS FOR RECOGNITION TASK 

You will now be seeing the names of several brands of 
candy bars. Half of these are brands that you have just 
received information on. Additionally, half of the list 
consists of brand names that you did not encounter before. 
Your task is to identify which are the brand names for which 
you have just seen some information. If you recognize a 
particular brand as one on which you have just seen some 
information, type "Y", if not, type "N" and press <RETURN>. 

Press <RETURN> to continue.... 



147 



APPENDIX IA-4 
LIST OF BRANDS USED AS DISTRACTORS IN RECOGNITION TASK 

Wunderbar 
Aero 
Coconut 
Bounty 
Glossettes 
Five Star 
Gems 

Jersey Milk 
Big Turk 
Nutties 
Snack 
Grand Slam 



148 



APPENDIX IB-1 
INSTRUCTIONS FOR BRAND NAME RECALL TASK 

Earlier you saw information on several brands of candy 
bars. Your task is to recall as many brand names as 
possible. Please list them in the space provided below. 



149 



APPENDIX IB-2 
INSTRUCTIONS FOR ATTRIBUTE RECALL TASK 

Earlier you saw information on several brands of 

Canadian candy bars . Your task is to recall as much of the 

information on what it contains. Write down what you can 
remember in as much detail as you can. 

BRAND NAME CONTAINS 

Crispy Crunch 



Caramilk 

Gold 

Smarties 

Breakaway 

Mirage 

Eat -more 

Mr. Big 

Crunchie 

Coffee Crisp 

Dairy Milk 

Sweet Marie 



150 



APPENDIX IB-3 
INSTRUCTIONS FOR PREFERENCE RATING TASK 

Listed below are twelve different brands of candy bars. 
Your task is to rate them on a scale of 1 (Extremely low in 
preference) to 7 (Extremely high in preference) . 

BRAND NAME PREFERENCE RATING 

Crispy Crunch 

Caramilk 

Gold 

Smarties 

Breakaway 

Mirage 

Eat-more 
Mr. Big 

Crunchie 

Coffee Crisp 

Dairy Milk 

Sweet Marie 



151 



APPENDIX IB-4 
INSTRUCTIONS FOR MOST LIKELY PRICE ESTIMATE 

"SKOR" is a brand of candy currently available in the 
Gainesville market. It is priced at 55 cents. If the new 
brands were also available, please try to estimate the most 
likely price at which you might expect to see them in most 
stores. Your task is to estimate the most likely price for 
each brand listed below. 

BRAND NAME MOST LIKELY PRICE 

Crispy Crunch 

Caramilk 

Gold 

Smarties 

Breakaway 

Mirage 

Eat-more 

Mr. Big 

Crunchie 

Coffee Crisp ^ 

Dairy Milk 
Sweet Marie 



152 



APPENDIX IB-5 
INSTRUCTIONS FOR HIGHEST (LOWEST) LIKELY PRICE ESTIMATES 

"SKOR" is a brand of candy currently available in the 
Gainesville market. It is priced at 55 cents in most stores. 
If the new brands were available in the market and you were 
to visit several stores, try to estimate the highest (lowest) 
likely price for these brands across stores. Your task is to 
estimate the highest (lowest) likely price for each brand 
listed below. 

BRAND NAME HIGHEST LIKELY LOWEST LIKELY 

PRICE PRICE 

Crispy Crunch 

Caramilk 

Gold 

Smarties 

Breakaway 

Mirage 

Eat-more 

Mr. Big 

Crunchie 

Coffee Crisp 

Dairy Milk 

Sweet Marie 



153 



APPENDIX IB-6 
INSTRUCTIONS FOR SIMILARITY RATING TASK 

Your task is to make similarity judgments between pairs 
of candy bars. The referent candy bar is SMARTIES. Please 
rate the similarity of this brand with other brands on a 
scale of 1 (Highly Dissimilar) to 5 (Highly Similar) . 

SMARTIES 





Highly 
Dissimil 


ar 






Highly 
Similar 


Crispy Crunch 


1 




2 


3 


4 


5 


Caramilk 


1 




2 


3 


4 


5 


Gold 


1 




2 


3 


4 


5 


Breakaway 


1 




2 


3 


4 


5 


Mirage 


1 




2 


3 


4 


5 


Eat-more 


1 




2 


3 


4 


5 


Mr. Big 


1 




2 


3 


4 


5 


Crunchie 


1 




2 


3 


4 


5 


Coffee Crisp 


1 




2 


3 


4 


5 


Dairy Milk 


1 




2 


3 


4 


5 


Sweet Marie 


1 




2 


3 


4 


5 



154 



APPENDIX IC-1 
INTRODUCTION BEFORE EXPOSURE TO ADVERTISING 

You have just been exposed to some information on 
several brand of candy bars. You will now be seeing ads for 
these brands before continuing to the next screen. 

The experimenter will show you the ads only when all of 
you are ready. So if you have finished before the others, 
please wait quietly for the other participants to finish. DO 
NOT PRESS <RETURN> NOW. 

Only after you have seen the ads should you press 
<RETURN> to continue. . . 



155 



APPENDIX IC-2 
INTRODUCTION TO MEMORIZATION TASK 

You will now be given a list consisting of brand names 
of candy bars available in Gainesville. You are required to 

memorize these names. 



156 



APPENDIX IC-3 
INSTRUCTIONS AND STIMULI FOR THE MEMORIZATION TASK 

Please take the next three minutes to memorize the 
following list of candy bar names: 

l.Cellas 2. Crunch 

3 . Whatchmacallit 4. Cote d'Or 

5. Pastilles Almond 6. Marabou 

V.Mars S.Twix 

9.Krackel lO.Kitkat 

ll.Tobler ll.Skor 

13. Almond Joy 12.100 Grand 

IS.Barnone 16 . Solitaires 

IV.Suvretti 18. M & M's 

19. Capers 20.Ferrara 

21. Mr. Goodbar 22 . Chocoletti 

23.Chocolat Lindt 24.Alpin Milch 



APPENDIX II 



DETAILED LISTING OF STIMULI AND TASKS FOR EXPERIMENT 3 



APPENDIX IIA-1 

PRICE PROFILES FOR THE TWELVE BRANDS 
ACROSS FIRST 8 SHOPPING TRIPS 









SHOPPING 


TRIPS 








1 


2 


3 


4 


5 


6 


7 


8 


BRAND 


















1 


L 


L 


L 


L 


L 


L 


L 


L 


2 


L 


L 


h 


L 


H 


H 


H 


H 


3 


L 


L 


L 


L 


H 


H 


H 


H 


4 


L 


L 


H 


H 


L 


L 


H 


H 


5 


L 


H 


L 


H 


L 


H 


L 


H 


6 


L 


L 


H 


H 


L 


L 


H 


H 


7 


L 


H 


H 


L 


H 


L 


L 


H 


8 


L 


H 


H 


L 


L 


H 


H 


L 


9 


L 


H 


L 


H 


H 


L 


H 


L 


10 


L 


H 


L 


H 


L 


H 


L 


H 


11 


L 


H 


H 


L 


H 


L 


L 


H 


12 


L 


L 


H 


H 


H 


H 


L 


L 



158 



159 



APPENDIX IIA-1 — continued 

PRICE PROFILES FOR THE TWELVE BRANDS 
ACROSS LAST 8 SHOPPING TRIPS 



SHOPPING TRIPS 
10 11 12 13 14 15 16 



BRAND 


















1 


H 


H 


H 


H 


H 


H 


H 


H 


2 


L 


L 


L 


L 


H 


H 


H 


H 


3 


H 


H 


H 


H 


L 


L 


L 


L 


4, 


L 


L 


H 


H 


L 


L 


H 


H 


5 


L 


H 


L 


H 


L 


H 


L 


H 


6 


H 


H 


L, 


L 


H 


H 


L 


L 


7 


H 


L 


L 


H 


L 


H 


H 


L 


8 


H 


L 


L 


H 


H 


L 


L 


H 


9 


H 


L 


H 


L 


L 


H 


L 


H 


10 


H 


L 


H 


L 


H 


L 


H 


L 


11 


L 


H 


H 


L 


H 


L 


L 


H 


12 


H 


H 


L 


L 


L 


L 


H 


H 



160 



APPENDIX IIA-2 



HIGHEST AND LOWEST PRICES FOR TWELVE BRANDS 
IN THE NO ADVERTISING CONDITION ^ 



BRAND 


HIGHEST LIKELY 


LOWEST LIKELY 




PRICE 


PRICE 


Crispy Crunch 


64.25 


41.75 


Caramilk 


63.63 


42.38 


Gold 


64.25 


41.13 


Smart ies 


60.50 


37.38 


Breakaway 


61.76 


39.88 


Mirage 


66.13 


43.00 


Eat-more 


60.50 


39.25 


Mr. Big 


72.37 


47.38 


Crunchie 


64.25 


44.25 


Coffee Crisp 


62.38 


41.75 


Dairy Milk 


64.88 


44.25 


Sweet Marie 


66.75 


46.12 



Mean 64.30 42.42 



^: Prices are in cent; 



161 



APPENDIX IIA-3 

HIGH AND LOW PRICES USED FOR THE EXPERIMENT 

BRAND HIGH PRICE LOW PRICE 

Crispy Crunch $0.65 $0.40 

Caramilk $0.55 $0.40 

Gold $0.65 $0.40 

Smarties $0.60 $0.35 

Breakaway $0.60 $0.35 

Mirage $0.65 $0.40 

Eat-more $0.60 $0.35 

Mr. Big $0.70 $0.45 

Crunchie $0.65 $0.40 

Coffee Crisp $0.60 $0.35 

Dairy Milk $0.65 $0.40 

Sweet Marie $0.70 $0.45 



162 



APPENDIX IIB-1 



BRAND-TRIPLES USED IN EXPERIMENT 3 



TRIPLE # 1 Crispy Crunch 

Caramilk 
Gold 

TRIPLE # 2 Smarties 

Breakaway 
Mirage 

TRIPLE # 3 Eat-more 

Mr. Big 
Crunchie 

TRIPLE # 4 Coffee Crisp 

Dairy Milk 
Sweet Marie 



163 



APPENDIX IIB-2 

GROUPING OF SEQUENCE FACTOR FOR COUNTERBALANCING ORDER 
OF PRESENTATION OF BRAND TRIPLES 





STIMULUS-BASED 


MEMORY-BASED 




NO-AD 


R 


D 


NO-AD 


R 


D 


ORDER 1 
SEQ 1,8,9,16 














ORDER 2, 
SEQ 3,6,11,14 














ORDER 3 
SEQ 2,7,10,15 














ORDER 4 
SEQ 4,5,12,13 















164 



APPENDIX IIB-3 



ORDER OF PRESENTATION OF PRODUCT AND PACKAGE INFORMATION 



ORDER 1 



ORDER 2 



ORDER 3 



ORDER 4 



POSITION 

2 3 4 



165 



APPENDIX IIB-4 



ORDER OF PRESENTATION OF ADVERTISING INFORMATION 

DURING SESSION 1 



ORDER 1 



ORDER 2 



ORDER 3 



ORDER 4 



POSITION 

2 3 4 



166 



APPENDIX IIC-1 



INTRODUCTION TO SECOND SESSION 



Welcome to the second part of the study on consumer 
responses to new brands of chocolate candy bars being 
considered for introduction into the Gainesville market. 

In this session you will be shopping for candy bars on 
the computer. As in the previous session, you will be 
required to use the keyboard to type in your responses, when 
prompted by messages on the screen. Please pay careful 
attention to these messages. If you have any questions, 
raise your hand and let the experimenter know that you need a 
clarification . 



Press <RETURN> to continue 



167 



APPENDIX IIC-2 



PRELIMINARY INSTRUCTIONS FOR SECOND SESSION 



Earlier in this experiment you have been exposed to 
information (brand name, weight and contents) on several 
brands of chocolate candy bars which are currently available 
in Canada, and which are now being considered for 
introduction into the Gainesville market . Imagine you are 
shopping for candy bars. You have $3.00 to spend on any of 
the candy bars that you have seen information on. If you are 
interested in getting price information on any of the brands, 
you will have to type in the name of the candy bar. You will 
be going on 15 of these shopping trips. Your purchases on 
any ONE of these trips will be randomly selected to decide 
which of the candy bars you will get . 

Remember, prices will be varying across these shopping 
trips, so that the best buy(s) on one particular trip need 
not be the same for all occasions. Thus, if you are to get 
your money's worth, you should be making reasoned choices. 

Press <RETURN> to continue.... 



168 



APPENDIX IIC-3 



INTRODUCTION TO THE PRACTICE SHOPPING TRIPS 



Before shopping for candy bars, you will go on four 
practice shopping trips to familiarize yourself with the 
shopping task on the computer. For the practice shopping 
trips, the product category will be soaps. Remember, the 
prices of the different brands will be varying from one 
shopping occasion to the next. 

The objective of the practice trips is to give you an 
idea of shopping on the computer. You can spend the $3.00 on 
one or more brands, in any combination you like. 

Press <RETURN> to continue.... 



169 



APPENDIX IIC-4 



INSTRUCTIONS FOR THE PRACTICE SHOPPING TRIPS 



Given below is a listing of all brands of soap that are 
available. If you would like price information on a 
particular brand, type in the brand name. You may do so for 
any number of brands on which you want price information, 
before making your choice. You may make a note of any 
information that the computer makes available. 

The following brands are available: 

Ivory 

Jergens 

Coast 

Shield 

Irish spring 

Dial 



TYPE IN THE BRAND NAME AND PRESS ENTER 



170 



APPENDIX IIC-4 — continued 



PRACTICE SHOPPING TRIP # 

You have $3.00 left to spend. 

Do you want to buy any more 
Which brand do you want to buy? 



BRAND NAME AMOUNT PURCHASED 

Ivory 

Jergens 

Coast 

Shield 

Irish Spring 

Dial 



171 



APPENDIX IID-1 



INSTRUCTIONS BEFORE EXPOSURE TO ADVERTISING 



You will now be seeing ads for the candy bars before you 
resume shopping. The experimenter will show the ads only 
when all of you are done with your shopping. So if you have 
finished before the others, please wait quietly for the other 
participants to finish also. DO NOT PRESS <RETURN> NOW. 



Only after you have watched the ads should you press 
<RETURN> to continue. . . . 



172 



APPENDIX IID-2 



INSTRUCTIONS FOR SHOPPING TRIPS — 
STIMULUS-BASED CONDITION 



After these practice shopping trips, now you will be 
shopping for candy bars. These shopping trips will be 
essentially similar to the practice trips. You are given a 
list of candy bars that are available and you have to type in 
the names of the candy bars on which you want to see price 
information. 

Remember, prices for these candy bars will be varying 
across purchase occasions, so that the best buys(s) on any 
particular trip need not be the same on another purchase 
occasion. Thus, if you are to get the best value for money, 
it is important that you make reasoned choices . Also, you 
are required to spend as much of your budget of $3.00 as you 
possible can. 

Press <RETURN> to continue.... 



173 



APPENDIX IID-2 — continued 



Given below is a listing of all brands of candy that are 
available. Please type in the name(s) of the candy bars that 
you want price information on. You may make a note of any 
information that the computer makes available. 



The following brands are available 

Crispy Crunch 
Caramilk 
Gold 
Smarties 
Breakaway 
Mirage 
Eat -more 
Mr. Big 
Crunchie 
Coffee Crisp 
Dairy Milk 
Sweet Marie 



TYPE IN THE BRAND NAME AND PRESS ENTER 



174 



APPENDIX IID-3 



INSTRUCTIONS FOR SHOPPING TRIPS 
MEMORY-BASED CONDITION 



After these practice shopping trips, now you will be 
shopping for candy bars. These are similar to the practice 
shopping trips except for one fact . In the practice trips 
you were given a list to decide the brands for which you 
wanted price information. But when you go on these trips, 
you will not have such a list of brand names available to 
you. You will have to retrieve from memory and type in the 
brand names that you want price information on. 

Remember, prices for these candy bars will be varying 
across purchase occasions, so that the best buys (s) on any 
particular trip need not be the same on another purchase 
occasion. Thus, if you are to get the best value for money, 
it is important that you make reasoned choices. Also, you 
are required to spend as much of your budget of $3.00 as you 
possible can. 

Press <RETURN> to continue.... 



175 



APPENDIX IID-3 — continued 



In the previous session you were exposed to several 
brand names (as well as weight and contents) of candy bars 
which are currently available in Canada, and which are now 
being considered for introduction into the Gainesville 
market. Please type in the name of the brand (s) of candy 
bars for which you want price information. You may make a 
note of any information that the computer makes available. 



TYPE IN THE BRAND NAME AND PRESS RETURN 



176 



APPENDIX HE 



ERROR TERMS USED TO TEST HYPOTHESIZED EFFECTS 
BRANDS ARE TREATED AS RANDOM 



QUASI F RATIOS 



MAIN EFFECT OF ADVERTISING: 



p , MS AD 



MSaDXBRAND + MSSUB/ENVXADXSEQ " MSbRANDXSUB/ENVXADXSEQ 



MAIN EFFECT OF DECISION ENVIRONMENT; 



^ msenv 



MSenVXBRAND + MSsUB/ENVXADXSEQ " MSbrANDXSUB/ENVXADXSEQ 



INTERACTION OF ENVIRONMENT AND ADVERTISING: 



^ mSenvxad 

msadxenvxbr + mssub/envxadxbr - msbrxsub/envxadxseq 



REFERENCES 



Alba, J. W., & A. Chattyopadhyay (1985), "The Effects of 
Part-List Cuing on Attribute Recall : Problem Framing at the 
Point of Retrieval," in Advances in Consumer Research , 12, 
eds . E. C. Hirschman & M. B. Holbrook, Provo, UT : Association 
for Consumer Research, 410-413. 

Alba, J. W., & A. Chattyopadhyay (1986), "Salience Effects in 
Brand Recall," Journal of Marketing Research , 23, 363-369. 

Alba, J. W., & J. W. Hutchinson (1990), "Public Policy 
Implications of Consumer Knowledge, " in Advances in Marketing 
and Public Policy , 2, ed. P. N. Bloom, Greenwich, CT : JAI 
Press, Inc, forthcoming. 

Alba, J. W., Hutchinson, J. W. , & J. G. Lynch ( 1990) , "Memory 
and Decision Making, " in Handbook of Consumer Theory and 
Research ^ eds. T. S. Robertson & H. K. Kassarjian, Prentice- 
Hall, Inc, forthcoming. 

Alba, J. W., & H. M. Marmorstein (1987), "The Effects of 
Frequency Knowledge on Consumer Decision Making, " Journal of 
Consumer Research , 14, 14-25. 

Albion, M. S., & P. W. Farris (1981), The Advertising 
Controversy , Boston, MA: Auburn House Publishing Company. 

Albion, M. S., & P. W. Farris (1983), Advertising's Hidden 
Effects: Manufacturers' Advertising and Retail Pricing , 
Boston, MA: Auburn House Publishing Company. 

Bain, J. S. (1956), Barriers to New Competition. Cambridge, 
MA: Harvard University Press. 

Benham, L.(1972), "The Effect of Advertising on the Price of 
Eyeglasses," Journal of Law and Economics ,, 15, 337-351. 

Bettman, J. R., & P. Kakkar (1977), "Effects of Information 
Presentation Format on Consumer Information Acquisition 
Strategies," Journal of Consumer Research , 3, 233-240. 

Bettman, J. R., & M. Sujan (1987), "The Effects of Framing on 
Evaluation of Comparable and Noncomparable Alternatives by 
Expert and Novice Consumers, " Journal of Consumer Research , 
14, 141-154. 

177 



178 

Biehal, G., & D. Chakravarti (1983), "Information 
Accessibility as a Moderator of Consumer Choice, " Journal of 
Consumer Research , 10, 1-14. 

Biehal, G., & D. Chakravarti (1986), "Consumers' Use of 
Memory and External Information in Choice: Macro and Micro 
Perspectives," Journal of Consumer Research , 12, 382-405. 

Bloom, P. N. (1976), Advertising^ Competition and Public 
Policy: A Simulation Study , Cambridge, MA: Ballinger 
Publishing Company. 

Boyd, H. W., Ray, M. L., & E. C. Strong (1972), "An 
Attitudinal Framework for Advertising Strategy, " Journal of 
Marketing , 36, 27-33. 

Cady, J, (1970), Restricted Advertising and Competition: The 
Case of Retail Drugs , Washington DC: American Enterprise 
Institute . 

Campbell, B. M. (1969), "The Existence and Determinants of 
Evoked Set in Brand Choice Behavior," Ph.D. dissertation, 
Columbia University. 

Chamberlain, E. H. (1933), The Theory of Monopolistic 
Competition , Cambridge, MA: Harvard University Press. 

Comanor, W. S., & T. A. Wilson (1974), Advertising and Market 
Power , Cambridge, Mass: Harvard University Press. 

Comanor, W. S., & T. A. Wilson (1979), "The Effects of 
Advertising on Competition: A Survey," Journal of Economic 
Literature , 17, 453-476. 

Cortsjens, M.L., & D. A. Gautschi (1983), "Formal Choice 
Models in Marketing," Marketing Science , 2, 19-56. 

Dorfman, R., & P. 0. Steiner (1954), "Optimal Advertising and 
Optimal Quality," American Economic Review , 44, 826-835. 

Eskin, G. J. (1975), "A Case for Test Marketing Experiments," 
Journal of Advertising Research , 15, 27-33. 

Eskin, G. J., & P. H. Baron (1977), "Effects of Price and 
Advertising in Test-Marketing Experiments," Journal of 
Marketing Research , 14, 499-508. 

Farris, P. W., & M. S. Albion (1980), "The Impact of 
Advertising on the Price of Consumer Products," Journal of 
Marketing , 44, 17-35. 



t 179 



Feldman, J. M., & J. G. Lynch (1988), "Self -Generated 
Validity and Other Effects of Measurement on Belief, 
Attitudes, Intentions and Behavior," Journal of Applied 
Psychology , 73, 421-435. 

Gardner, M. P. (1983), "Advertising Effects on Attributes 
Recalled and Criteria Used for Brand Evaluations," Journal of 
Consumer Research , 10, 310-318. 

Hauser, J. R., & b. Wernerfelt (1990), "An Evaluation Cost 
Model of Evoked Sets," Journal of Consumer Research , 16, 393- 
408. 

Hoch, S. J. & Deighton, J. (1989), "Managing What Consumers 
Learn from Experience," Journal of Marketing , 53, 1-20. 

Hoch, S. J. & Y. Ha (1986), "Consumer learning. Advertising 
and the Ambiguity of Product Experience, " Journal of Consumer 
Research , 13, 221-233. 

Howard, J. A., & J. N. Sheth (1969), The Theory of Buyer 
Behavior , New York: John Wiley. 

Hoyer, W. D. (1984), "An Examination of Consumer Decision 
Making for a Common Repeat Purchase Product," Journal of 
Consumer Research , 11, 822-829. 

Huber, J., Holbrook, M. B., & B. Kahn (1986), "The Effects of 
Competitive Context and Additional Information on Price 
Sensitivity," Journal of Marketing Research , 23, 250-260. 

Hutchinson, J. W. (1986), "Discrete Attribute Models of Brand 
Switching," Marketi ng Science , 5, 350-371. 

Johnson, E. J., & r. j. Meyer (1984), "Compensatory Choice 
Models of Noncompensatory Processes: The Effect of Varying 
Context," Journal of Consumer Research , 11, 528-541. 

Kaldor, N. (1950), "The Economic Aspects of Advertising," 
Review of Economic Studies , 18, 1-27. 

Keller, K. L. (1987), "Memory in Advertising: The Effect of 
Advertising Memory Cues on Brand Evaluations," Journal of 
Consumer Research , 14, 316-333. 

Keppel, G. (1982), Design and Analvsis: A Researcher's 
Handbook, Englewood Cliffs, NJ: Prentice-Hall, Inc. 



180 

Kirmani, A., & P. Wright (1989), "Money Talks: Perceived 
Advertising Expense and Expected Product Quality, " Journal of 
Consumer Research , 16, 344-353. 

Krishnamurthi, L., & S. P. Raj (1985), "The Effect of 
Advertising on Consumer Price Sensitivity, " Journal of 
Marketing Research , 22, 119-129. 

Lambin, J. J. (1976), Advertising, Competition and Market 
Conduct in Oligopoly over Time , Amsterdam: North-Holland 
Publishing Co. 

Leffler, K. B. (1982), "Persuasion or Information? The 
Economics of Prescription Drug Advertising, " Journal of Law 
and Economics , 24, 45-74. 

Lutz, R. J. (1975), "Changing Brand Attitudes through 
Modification of Cognitive Structure, " Journal of Consumer 
Researc h, 1, 49-59. 

Lynch, J. G., & P. N. Bloom (1987), "Psychological 
Perspectives on the Economics of Advertising," Paper 
presented at the Annual Conference of the Association for 
Consumer Research, Cambridge, MA. 

Lynch, J. G., Chakravarti, D., & Mitra A. (1990), 
"Distinguishing Contrast Effects Caused by Changes in Mental 
Representation from those Caused by Changes in Response 
Language ,. Working Paper, Center for Consumer Research, 
University of Florida, Gainesville. 

Lynch, J. G., Marmorstein, H. M. & M. F. Weigold (1988), 

"Choices from sets Including Remembered Brands: Use of 

Recalled Attributes and Prior Overall Evaluations, " Journal 
of Consumer Research , 15, 225-233. 

Lynch, J. G., & T. R. Srull (1982), "Memory and Attentional 
Factors in Consumer Choice: Concepts and Research Methods," 
Journal of Consumer Research , 9, 18-37 . 

McAulife, R. E. (1987), Advertising, Competition and Public 
Policy: Theories and New Evidence ^ Lexington MA: D. C. Heath 
& Co. 

Milgrom, P., & J. Roberts (1986), "Price and Advertising as 
Signals of Quality," Journal of Political Economy , 94, 796- 
821. 

Moran, W. T. (1978), Information for Advertising Budgeting , " 
(report), Greenwich, CT : Moran Inc. 



181 



Narayana C. L., & R. J. Markin (1975), "Consumer Behavior and 
Product Performance: An Alternative Conceptualization," 
Journal of Marketing , 39, 1-6. 

Nedungadi, P. (1987), "Formation and Use of a Consideration 
Set," Ph.D. dissertation. University of Florida. 

Nedungadi, P., & J. W. Hutchinson (1985), "The 
Protot ypicality of Brands: Relationships with Brand 
Awareness, Preference and Usage," in Advances in Consumer 
Research , 12, eds . E. C. Hirschman & M. B. Holbrook, Provo, 
UT : Association for Consumer Research, 498-503. 

Nelson, P. H. (1970), "Information and Consumer Behavior," 
Journal of Political Economy , 78, 311-329. 

Nelson, P. H. (1974a), "Advertising as Information," Journal 
of Political Economy , 81, 729-745. 

Nelson, P. H. (1974b), "The Economic Value of Advertising," 
in Advertising and Society , ed. Y. Brozen, New York: New York 
University Press, 43-65 

Nelson, P. H. (1975), "The Economic Consequences of 
Advertising," Journal of Business , 48, 213-241. 

Nelson, P. H. (1978), "Advertising as Information Once More," 
in Issues in Advertising: The Economics of Persuasion , ed. D. 
G. Tuerck, New York: New York University Press, 133-161. 

Nerlove, M., & K. J. Arrow (1962), "Optimal Advertising 
Policy under Dynamic Conditions," Economica , 29, 129-142. 

Norris, V. P. (1984), "The Economic Effects of Advertising: A 
Review of the Literature, " Current Issues and Research in 
Advertising , 39-134. 

Ornstein, S. I. (1977), Industrial Competition and 
Advertising Intensity , Washington DC: American Enterprise 
Institute . 

Park. C. W. Iyer, E. S., & D. C. Smith (1989), "The Effects 
of Situational Factors on In-Store Grocery Shopping Behavior: 
The Role of Store Environment and Time Available for 
Shopping," Journal of Consumer Research , 15, 422-433. 

Parkinson T. L., & M. Reilly (1979), "An Information 
Processing Approach to Evoked Set Formation, " in Advances in 



182 

Consumer Research ^ 6, ed . W. W. Wilkie, Ann Arbor, MI: 
Association of Consumer Research, 227-231. 

Prasad V. K., & L. W. Ring (1976), "Measuring Sales Effects 
of Some Marketing Mix Variables and their Interactions," 
Journal of Marketing Research , 13, 391-396. 

Shugan, S. M. (1980), "The Cost of Thinking," Journal of 
Consumer Research , 7, 99-111. 

Satterthwaite, F. E. (1946), "An Approximate Distribution of 
Estimates of Variance Components," Biometrics Bulletin, 2, 
110-114. 

Silk, A. J., & G. L. Urban (1978), "Pre-Test Market 
Evaluation of New Packaged Goods: A Model and Measurement 
Methodology," Journal of Marketing Research , 15, 171-191. 

Steiner, R. L. (1978), "A Dual Stage Approach to the Effects 
of Advertising on Competition and Prices, in Marketing and 
the Public Interest , Report No. 78-105, ed. J. Cady, Boston, 
MA: Marketing Science Institute, 127-150. 

Stigler, G. (1961), "The Economics of Information," Journal 
of Political Economy , 69, 213-225. 

Telser, L. G. (1963), "Advertising and Competition," Journal 
of Political Economy , 72, 537-562. 

Upshaw, H. S. (1962), "Own Attitude as an Anchor in Equal 
Appearing Intervals," Journal of abnormal and Social 
Psychology , 64, 85-96. 

Wyer, R. S. (1974), Cognitive Organization and Change: An 
Information Processing Approach , Potomac, MD : Lawrence 
Earlbaum Associates. 

Wilkie, W. W., & P. W. Farris (1976), Consumer Information 
Processing: Perspectives and Implications for Advertising , 
Report No. 76-113, Cambridge, MA: Marketing Science 
Institute . 

Wilkie, W. W., & E. A. Pessemier (1973), "Issues in 
Marketing's use of Multi-Attribute Models," Journal of 
Marketing Research , 10, 428-441. 

Wittink, D. R. (1977), "Advertising Increases Sensitivity to 
Price," Journal of Advertising Research , 17, 39-42. 



BIOGRAPHICAL SKETCH 

Anusree was born January 10,1961, in Calcutta, India. 
From 1955 to 1976 she attended Modern High School at Calcutta 
and completed the Higher Secondary Examination in 1979. For 
the next three years she pursued an undergraduate degree in 
economics at the Presidency College at Calcutta. She 
graduated in 1982 with honors in economics and a minor in 
mathematics . 

She then entered the Post-graduate Program in Management 
at the Indian Institute of Management (I.I.M.) Calcutta in 
July 1982. In April 1984, she received her master's degree 
in Management from I.I.M. , having specialized in marketing 
and finance. 

In 1984 Anusree joined Lintas India Limited, a member of 
the SSC&B: LINTAS Worldwide Advertising Network. She worked 
as account executive and worked on advertising campaigns for 
Lever Brothers, Union Carbide, among other accounts. 

She gave up the job in advertising in order to pursue 
doctoral studies in the United States. In Fall 1985, she was 
admitted to the Ph.D. program in marketing at the University 
of Florida and has worked towards her degree since that time. 

During her term in the doctoral program, Anusree 
acquired teaching experience as an instructor for the 



183 



184 



undergraduate marketing research course. She also assisted 
with various undergraduate and graduate marketing classes. 
Additionally, she has worked on several academic consumer 
behavior research projects. 

She accepted an offer for a position as Assistant 
Professor of Marketing at the American University. In summer 
1990 she assumed the full-time position. 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degree of Doctor of Philosophy. 




n 




Johyl/ G. Lynch, Chairrdati ^y 
Associate Professor of Marketing 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degree of Doctor of Philosophy. 




Associate Professor of Marketing 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degree o//f)g^torybfy^ilosophy . 



Ja 



JX Wesle^Hxitchinson 
■^ssocijxe Professor of Marketing 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degr^^e of Doctor of ^ilosophy. 



//'OO- 




Bartonv A. Weitz 
Professor of Marketing 



I certify that I have read this study and that in my 
opinion it conforms to acceptable standards of scholarly 
presentation and is fully adequate, in scope and quality, as 
a dissertation for the degree of Doctor of Philosophy. 



Richard E . Romano 

Associate Professor of Economics 



This dissertation was submitted to the Graduate Faculty 
of the Department of Marketing in the College of Business 
Administration and to the Graduate School and was accepted as 
partial fulfillment of the requirements for the degree of 
Doctor of Philosophy. 



August 19 90 



Dean, Graduate School 






(.